0% found this document useful (0 votes)

92 views137 pages

FusionCompute V100R005C10 Host and Cluster Management Guide 01

Uploaded by

Alex

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

92 views137 pages

FusionCompute V100R005C10 Host and Cluster Management Guide 01

Uploaded by

Alex

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 137

FusionCompute

V100R005C10

Host and Cluster Management

Guide

Issue 01
Date 2015-11-11

HUAWEI TECHNOLOGIES CO., LTD.

Copyright © Huawei Technologies Co., Ltd. 2015. All rights reserved.
No part of this document may be reproduced or transmitted in any form or by any means without prior written
consent of Huawei Technologies Co., Ltd.

Trademarks and Permissions

and other Huawei trademarks are trademarks of Huawei Technologies Co., Ltd.
All other trademarks and trade names mentioned in this document are the property of their respective
holders.

Notice
The purchased products, services and features are stipulated by the contract made between Huawei and the
customer. All or part of the products, services and features described in this document may not be within the
purchase scope or the usage scope. Unless otherwise specified in the contract, all statements, information,
and recommendations in this document are provided "AS IS" without warranties, guarantees or
representations of any kind, either express or implied.

The information in this document is subject to change without notice. Every effort has been made in the
preparation of this document to ensure accuracy of the contents, but all statements, information, and
recommendations in this document do not constitute a warranty of any kind, express or implied.

Huawei Technologies Co., Ltd.

Address: Huawei Industrial Base
Bantian, Longgang
Shenzhen 518129
People's Republic of China

Website: http://e.huawei.com

Issue 01 (2015-11-11) Huawei Proprietary and Confidential i

About This Document

Purpose
This document describes how to create, configure, adjust, and reclaim host and cluster
resources on the FusionCompute.

Intended Audience
This document is intended for:
l Technical support engineers
l Maintenance engineers

Symbol Conventions
The symbols that may be found in this document are defined as follows:

Symbol Description

Indicates an imminently hazardous situation which, if

not avoided, will result in death or serious injury.
DANGER

Indicates a potentially hazardous situation which, if not

avoided, could result in death or serious injury.
WARNING

Indicates a potentially hazardous situation which, if not

avoided, may result in minor or moderate injury.
CAUTION

Indicates a potentially hazardous situation which, if not

avoided, could result in equipment damage, data loss,
NOTICE
performance deterioration, or unanticipated results.
NOTICE is used to address practices not related to
personal injury.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential ii

Symbol Description

NOTE Calls attention to important information, best practices

and tips.
NOTE is used to address information not related to
personal injury, equipment damage, and environment
deterioration.

GUI Image Conventions

The GUI images that may be found in this document are for reference only. See the product
GUI in practice.

Change History
Changes between document issues are cumulative. The latest document issue contains all the
changes made in earlier issues.

Issue 01 (2015-11-11)
This issue is the first official release.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential iii

Contents

About This Document.....................................................................................................................ii

1 Overview of Host and Cluster Management........................................................................... 1
2 Cluster Management.....................................................................................................................5
2.1 Creating a Cluster........................................................................................................................................................... 6
2.2 Querying Cluster Information.......................................................................................................................................11
2.3 Modifying the Name and Description of a Cluster.......................................................................................................12
2.4 Managing Clusters Using Cluster Folders....................................................................................................................13
2.5 Configuring the Computing Resource Scheduling Policies......................................................................................... 14
2.6 Configuring the HA Policy for a Cluster......................................................................................................................23
2.7 Configuring the VM Processing Policy upon Data Store Faults.................................................................................. 26
2.8 Enabling Memory Overcommitment in a Cluster........................................................................................................ 27
2.9 Configuring the VM Startup Policy..............................................................................................................................29
2.10 Configuring Guest NUMA for a Cluster.................................................................................................................... 30
2.11 Configuring IMC for a Cluster................................................................................................................................... 32
2.12 Removing a Cluster.................................................................................................................................................... 34

3 Host Management....................................................................................................................... 35
3.1 Adding Hosts................................................................................................................................................................ 36
3.2 Querying Host Information.......................................................................................................................................... 39
3.3 Modifying the Name and Description of a Host...........................................................................................................40
3.4 Modifying the Host name in OS...................................................................................................................................41
3.5 Changing the Switching Mode of a Network Port....................................................................................................... 42
3.6 Updating the Number of Virtual Ports of the SR-IOV-enabled Network Port............................................................. 43
3.7 Setting a Host to the Maintenance Mode......................................................................................................................44
3.8 Changing the Multipathing Type of a Host.................................................................................................................. 45
3.9 Setting Time Synchronization on a Host...................................................................................................................... 47
3.10 Performing Forcible Time Synchronization............................................................................................................... 49
3.11 Configuring Host BMC Parameters............................................................................................................................50
3.12 Sharing the GPU on a Host.........................................................................................................................................51
3.13 Configuring the Host GPU Mode............................................................................................................................... 52
3.14 Enabling or Disabling LLDP for a Host..................................................................................................................... 54
3.15 Enabling the Host Antivirus Function........................................................................................................................ 55
3.16 Moving a Host............................................................................................................................................................ 56

Issue 01 (2015-11-11) Huawei Proprietary and Confidential iv

3.17 Removing a Host........................................................................................................................................................ 58

4 System Port Management.......................................................................................................... 61

4.1 Binding Network Ports................................................................................................................................................. 62
4.2 Adding a Storage Port...................................................................................................................................................68
4.3 Adding Service Management Ports (Required Only for Virtualized SAN Storage).................................................... 74
4.4 Querying Routes of a System Port............................................................................................................................... 79
4.5 Modifying Attributes of a System Port.........................................................................................................................80
4.6 Deleting a Storage Port.................................................................................................................................................85
4.7 Deleting a Service Management Port........................................................................................................................... 85

A Appendix......................................................................................................................................87
A.1 FAQ..............................................................................................................................................................................88
A.1.1 How to Configure Static Routes for Storage Ports on a Host...................................................................................88
A.1.2 How to Identify Server Ports.................................................................................................................................... 90
A.1.3 Failed to Execute the Policy for Scheduling Computing Resources in a Cluster.....................................................95
A.1.4 Background Information...........................................................................................................................................96
A.1.5 Compatibility.......................................................................................................................................................... 101
A.2 Parameter Reference.................................................................................................................................................. 102
A.2.1 Cluster Parameters.................................................................................................................................................. 102
A.2.2 Computing Resource Scheduling Rules Parameters...............................................................................................107
A.2.3 HA Parameters........................................................................................................................................................ 113
A.2.4 VM Startup Policy Parameters................................................................................................................................116
A.2.5 Host Parameters...................................................................................................................................................... 116
A.2.6 Time Synchronization Parameters.......................................................................................................................... 117
A.2.7 BMC Parameters.....................................................................................................................................................118
A.2.8 Network Ports Parameters.......................................................................................................................................118
A.2.9 Storage Port Parameters..........................................................................................................................................123
A.2.10 System Port Parameters........................................................................................................................................ 125
A.3 Parameters in Advanced Settings for Computing Resource Scheduling...................................................................127

Issue 01 (2015-11-11) Huawei Proprietary and Confidential v

Copyright © Huawei Technologies Co., Ltd.
FusionCompute
Host and Cluster Management Guide 1 Overview of Host and Cluster Management

1 Overview of Host and Cluster Management

Summary
FusionCompute resources include host and cluster resources, network resources, and storage
resources. Host and cluster management involves the following operations for the
FusionCompute: create a cluster or host and adjust host or cluster configurations.

Working Principles of Computing Resource Virtualization

The FusionCompute system integrates physical CPUs and memory resources on hosts into a
computing resource pool and divides the resources into virtual CPUs and memory resources
for VMs, as shown in Figure 1-1.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 1

Copyright © Huawei Technologies Co., Ltd.
FusionCompute
Host and Cluster Management Guide 1 Overview of Host and Cluster Management

Figure 1-1 Working principles of computing resource virtualization

Virtual CPU
Virtual memory

Virtualized
computing
resource pool
Memory resource pool CPU resource pool

Virtualization
layer

Host
(physical server)

Table 1-1 describes the concepts in the preceding figure.

Table 1-1 Concepts in computing resource virtualization

Name Description

Host (physical After the FusionCompute host OS is installed on a physical server,

server) this physical server is capable of virtualizing computing resources
(CPUs and memory) on it, that is, generates a virtualization layer.
Virtualization layer

Virtualized Virtualized computing resources on hosts are integrated into a

computing resource computing resource pool. Typically, a computing resource pool
pool contains resources from all hosts in the same cluster.

Virtual CPU and When the system creates a VM, the system automatically allocates
memory required memory space and virtual CPUs from the resource pool to
the VM according to the specified VM specifications.
VM
NOTE
The virtual CPU and memory resources used by a VM must be provided by
the same host. If this host fails, the system automatically assigns another host
to the VM to provide computing resources. Therefore, the resources actually
used by a VM cannot exceed the specifications of the hardware resources on
the host.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 2

Copyright © Huawei Technologies Co., Ltd.
FusionCompute
Host and Cluster Management Guide 1 Overview of Host and Cluster Management

Resource Creation Process

Figure 1-2 shows the process for creating a host or cluster.

Figure 1-2 Creation process

Host and cluster
Storage management Network management
management
Local or raw
device mapping
Create a cluster. What is the FC SAN Yes
storage resource Is a DVS available?
type?

Add hosts to the cluster. IP SAN NAS No

FusionStorage
Create a DVS.
Adding storage resources Configure the FC SAN
Set host clock to the site. initiator.
synchronization.

Add an uplink.
Associate the storage
No resource with hosts.
Is a Huawei SAN
device or an FC
SAN device (Optional) Add a VLAN
used? pool.
Scan storage devices.
Yes
Change the multiple (Optional) Add a subnet.
storage path.
Add a data store.

Create a port group.

Is the host
connected to the No Create a disk.
storage device
using ISCSI.

Yes
Add a storage port to the
host.

Procedure
Table 1-2 describes the creation procedure.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 3

Copyright © Huawei Technologies Co., Ltd.
FusionCompute
Host and Cluster Management Guide 1 Overview of Host and Cluster Management

Table 1-2 Procedure

Resource Procedure Description

Host and 1. Creating a Cluster To adjust configurations

cluster 2. Adding Hosts of cluster after creating it,
see the following sections:
3. If an external clock source needs to be
configured for the host, perform operations l Modifying the Name
provided in Setting Time Synchronization and Description of a
on a Host. Cluster
4. If the in-use storage device is a Huawei l Configuring the
storage area network (SAN) device, perform Computing Resource
operations provided in Changing the Scheduling Policies
Multipathing Type of a Host. l Configuring the HA
5. If the host connects to the storage device Policy for a Cluster
using an Internet Small Computer Systems l Enabling Memory
Interface (iSCSI), perform operations Overcommitment in a
provided in Adding a Storage Port. Cluster
After host configuration, perform operations for l Configuring the VM
configuring the host store and network. Startup Policy
To adjust configurations
of a host after creating it,
see the following sections:
l Modifying the Name
and Description of a
Host
l Modifying Attributes
of a System Port

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 4

2 Cluster Management

About This Chapter

2.1 Creating a Cluster

2.2 Querying Cluster Information
2.3 Modifying the Name and Description of a Cluster
2.4 Managing Clusters Using Cluster Folders
2.5 Configuring the Computing Resource Scheduling Policies
2.6 Configuring the HA Policy for a Cluster
2.7 Configuring the VM Processing Policy upon Data Store Faults
2.8 Enabling Memory Overcommitment in a Cluster
2.9 Configuring the VM Startup Policy
2.10 Configuring Guest NUMA for a Cluster
2.11 Configuring IMC for a Cluster
2.12 Removing a Cluster

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 5

2.1 Creating a Cluster

Scenarios
On FusionCompute, create a user cluster based on the data plan.

Prerequisites
Conditions
You have logged in to FusionCompute.

Procedure
Switch to the page for creating a cluster.
1 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
2 In the navigation tree on the left, right-click the site and choose Create Cluster.
The Create Cluster page is displayed.
Configure the basic information.
3 Configure the name and description of the cluster.
4 Click Next.
The Basic Configuration page is displayed.
5 If you need to use the host memory overcommitment function, set Host memory
overcommitment to Enable.
If the host memory overcommitment function is enabled, the memory provided by the
host for VMs can be greater than the physical memory of the host, and therefore the VM
density supported by the host can be increased.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 6

NOTE

– After host memory overcommitment is enabled, the overcommitment ratio for a VM can be
controlled by specifying the Reserved (MB) parameter in the VM QoS settings area.
n If the configured VM memory is greater than or equal to 16 GB, the VM Reserved
(MB) parameter must be set to its maximum value to ensure the optimal VM
performance. In this case, host memory overcommitment does not take effect for the
VM.
n If the configured VM memory is less than 16 GB, the VM Reserved (MB) value can be
set to 70% of the specified VM memory size so that more VMs than the number of
physically supported ones can use the host memory resources in practice. If the
monitored VM memory usages have been persistently greater than 40% in several hours,
the VM Reserved (MB) parameter must be set to its maximum value. In this case, host
memory overcommitment does not take effect for the VM.
– After host memory overcommitment is enabled, the total available memory capacity equals to
the total memory capacity in the virtualization domain multiplied by 120%. The total memory
capacity in the virtualization domain equals to the server memory capacity minus memory size
required by virtualization management. You can choose Computing Pool > Site > Cluster >
Host > Summary > Monitoring Information to view the total memory capacity in the
virtualization domain.
– After host memory overcommitment is enabled, plan VMs on hosts based on the total memory
capacity. If VMs that consumed a large amount of memory exist, some VMs may fail to start
even if the used memory has been released. This startup failure occurs because the
virtualization layer does not know that the memory is released.
– After host memory overcommitment is enabled, VMs cannot be hibernated and memory
snapshots cannot be created for them.
6 Configure the VM startup policy.
– Assign automatically: The system starts a VM on a host that has available
resources in the cluster.
– Assign based on load balancing: The system starts VMs on the host with the
largest available CPU capacity.
7 If you need the guest non-uniform memory access (NUMA) function, set GuestNUMA
to Enable.
Guest NUMA presents a topology view of memory and CPU resources on each host to
VMs, the VM user can configure VM CPUs and memory using the third-party software
(such as Eclipse) based on this topology, so that VMs can obtain the most easy-to-access
memory based on the topology, thereby reducing access latency and improving VM
performance.
NOTE
Ensure that the following conditions are met for the Guest NUMA function to take effect:
– The number of VM CPUs supported by a host in the cluster must be a multiple of the physical
CPU quantity on the host or a multiple of the thread quantity of a single CPU of the host. (To
check the number of physical CPUs on a host or the number of CPU threads of a host, switch
to the host hardware information page and choose Hardware > CPU, as described in
Querying Host Information.)
The Guest NUMA function may fail on the host that accommodates the VM if the host CPU
quantity changes (due to VM live migration or VM startup on another host) or VM CPU
specifications change.
– The memory overcommitment function or the host CPU resource mode is disabled for the
cluster.
– The NUMA function is enabled for hosts in the cluster. For example, to enable the NUMA
function for a RH2288H V2 server, choose Advanced > Advanced Processor in the advanced
basic input/output system (BIOS) settings of the server, and set NUMA mode to Enabled.
– VMs on the hosts in the cluster are restarted after the Guest NUMA function is enabled for the
cluster.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 7

8 Set VMs processing policy upon data store faults as needed.

This parameter can be set to No processing or Stop VM.
9 If local RAM disks are required for hosts in the cluster, set Local Ramdisk to Enable.
10 Click Next.
The Configure HA page is displayed.
Configure high availability (HA) settings.
11 Determine whether to configure HA settings.

The HA function can be enabled for VMs in a cluster only when the cluster also has the
HA function enabled.

– If yes, go to 12.
– If no, click Next and go to 14.
12 Select Enable.
13 Configure the HA function.

The HA setting can be configured to HA resource reservation or Tolerate cluster host

failures.

– HA resource reservation: The system reserves the specified amount of CPU and
memory resources for the cluster. The reserved resources can only be used to
implement the VM HA function.
– Tolerate cluster host failures: The system allows the specified number of hosts to
become faulty in the cluster. The system also periodically checks whether sufficient
resources are available in the cluster to support service switchover for these faulty
VMs. If no sufficient resources are available, an alarm will be generated to notify
users of ensuring sufficient resources in the cluster for VM service switchover.
A slot is a basic unit for allocating CPU and memory resources, and the slot value
can be set to Automatic or Custom.
n Automatic: The system sets the slot size based on the maximum amount of
VM CPU and memory resources required by the cluster.
n Custom: Users can set custom VM CPU and memory resource size based on
the service requirements.
14 Click Next.
The Configure Resource page is displayed.
Configure computing resource scheduling policies.
15 Determine whether to enable the computing resource scheduling function.
– If yes, go to 16.
– If no, click Next and go to 27.
16 Select Enable computing resource scheduling.
17 Select an automation level.
Automation levels include:
– Manual: The user is prompted to migrate VMs based on the suggestions provided
on the Computing Resource Scheduling page.
– Automatic: The system automatically migrates VMs to maximize resource
utilization when the resource load is heavy.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 8

18 Select a measurement condition.

Measurement conditions (such as CPU or memory usage) are criteria for the system to
determine whether to schedule resources.
Measurement conditions include:
– CPU
– Memory
– CPU and Memory
19 Configure the CPU and memory thresholds for triggering scheduling policies.
20 Set resource scheduling thresholds for each hour.
Thresholds include:
– Conservative: The system does not migrate VMs or provide VM migration
suggestions.
– Slightly conservative, Medium, Slightly radical, and Radical: The system
improves cluster load imbalance by a degree in ascending order.
The threshold is set to Medium for each period by default.
The threshold validity period can be set to:
– By day: The thresholds always take effect.
– By week: The thresholds take effect only on the specified days in a week.
– By month: The thresholds take effect only on the specified days in a month.
21 Set the VM scheduling customization value to configure the VM scheduling
automation level.
– If this parameter is selected, you can set an automation level for each VM in the
cluster on the VM Override Policy page.
– If this parameter is not selected, the VMs in a cluster use the automation level
specified for the cluster.
22 Determine whether to enable automated power management.
– If yes, go to 23.
– If no, click Next and go to 27.
With this function enabled, the system migrates the VMs on the host and powers on or
off the host in the cluster based on the host resource usages. This function is remotely
implemented by the host BMC. Therefore, you need to configure the BMC for all hosts
in the cluster.
23 Select Enable automated power management.
Automated power management depends on automated computing resource scheduling.
Therefore, automated power management takes effects only when automated computing
resource scheduling is enabled and the migration threshold is not set to Conservative.
24 Configure the automation level.
Automation levels include:
– Manual: The user is prompted to migrate VMs based on the suggestions provided
on the Computing Resource Scheduling page.
– Automatic: The system automatically migrates VMs to maximize resource
utilization when the resource load is heavy.
25 Set the power management threshold and the threshold validity duration for each period.
Power management thresholds include:

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 9

– Conservative: The system does not power off hosts by default. It powers on
available hosts in the cluster only when the average host resource usage in the
cluster is higher than the heavy-load threshold.
– Slightly conservative, Medium and Slightly radical: The system powers on
available hosts in the cluster when the average host resource usage in the cluster is
higher than the heavy-load threshold. It powers off some hosts when the average
host resource usage in the cluster is lower than the light-load threshold.
– Radical: The system does not power on available hosts by default. It powers off
some hosts in the cluster only when the average host resource usage in the cluster is
lower than the light-load threshold.
The default value is Medium for all periods.

Table 2-1 lists the light-load and heavy-load threshold value of each power management
threshold.

Table 2-1 Light-load and heavy-load threshold value of each power management
threshold
Threshold Name Heavy-Load Threshold Light-Load Threshold
Value Value

Conservative 63% N/A

Slightly conservative 72% 23%

Medium 81% 45%

Slightly radical 90% 54%

Radical - 63%

The threshold validity period can be set to:

– By day: The thresholds always take effect.
– By week: The thresholds take effect only on the specified days in a week.
– By month: The thresholds take effect only on the specified days in a month.
26 Click Next.
The Configure IMC page is displayed.
Configure the incompatible migration cluster (IMC) function.
27 Determine whether to enable the IMC function.
The IMC function allows the VMs running on the hosts in the cluster to present the same
CPU features. This ensures successful VM migration across these hosts even if the hosts
use physical CPUs of different performance baselines.
The IMC mode of a cluster must be set to the CPU generation that exposes the minimum
function set in the cluster or an earlier version of the CPU generation.
The CPU generation of the host to be added to the IMC-enabled cluster is the same as or
later than the IMC mode configured for the cluster.
FusionCompute supports the following Intel CPU generations, which present higher
performance levels in ascending order:

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 10

– Merom
– Penryn
– Nehalem
– Westmere
– Sandy Bridge
– Ivy Bridge
NOTE
After setting an IMC mode for a cluster, you need to enable the Execute Disable Bit function,
which is also known as the No eXecute bit (NX) or eXecute Disable (ND) function, in the BIOS
advanced CPU options for existing hosts in the cluster and the hosts to be added to the cluster.
28 Set the IMC mode and description.
29 Click Next.
The Confirm page is displayed.
Finish cluster creation.
30 Click Create.
An information dialog box is displayed.
31 Click OK.
The cluster creation task is complete.
Follow-up Procedure
Add hosts to the cluster.
----End

2.2 Querying Cluster Information

Scenarios
On FusionCompute, query the information about a cluster.

Prerequisites
Conditions
You have logged in to FusionCompute.
Data
You have obtained the name of the cluster to be queried.

Procedure

1 Perform the following operations based on the information to be queried:

– To query the general information about the cluster, go to 2.
– To query the information about VMs in the cluster, go to 5.
– To query the information about devices used by the cluster, go to 8.
– To query the information about hosts in the cluster, go to 11.
Query the general information about the cluster.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 11

2 On FusionCompute, click Computing Pool.

The Computing Pool page is displayed.
3 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.
4 Query the general information about the cluster on the Summary page.
No further action is required.
Query the information about VMs in the cluster.
5 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
6 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.
7 Query the information about VMs in the cluster on the VM page.
No further action is required.
Query the information about devices used by the cluster.
8 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
9 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.
10 Query the information about devices used by the cluster on the Device page.
No further action is required.
Query the information about hosts in the cluster.
11 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
12 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.
13 Query the information about hosts in the cluster on the Host page.
No further action is required.

----End

2.3 Modifying the Name and Description of a Cluster

Scenarios
On FusionCompute, modify the name and description of a cluster.

Prerequisites
Conditions
You have logged in to FusionCompute.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 12

Procedure
Step 1 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
Step 2 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.

Step 3 In the Basic Information area on the Summary page, click in the Name and
Description rows.
A dialog box is displayed.
Step 4 Enter the new name and description of the cluster.

Step 5 Click OK.

The cluster information is modified.

----End

2.4 Managing Clusters Using Cluster Folders

Scenarios
On FusionCompute, create cluster folders, move existing clusters to or create clusters in the
cluster folders so that clusters running different services can be managed by folder.

Prerequisites
Conditions
You have logged in to FusionCompute.

Procedure
Create a cluster folder.
1 On FusionCompute, click Computing Pool.
2 In the navigation tree on the left, right-click the target cluster and choose Create Cluster
Folder.
3 Enter the cluster folder name and click OK.
An information dialog box is displayed.
4 Click OK.
The cluster folder is created.
Create a cluster in a cluster folder.
5 In the navigation tree, right-click the cluster folder and choose Create Cluster.
6 Set the cluster parameters as instructed.
For details about the parameters, see Creating a Cluster.
Move a cluster.
7 In the navigation tree, right-click the cluster and choose Move.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 13

8 Select the destination cluster folder or site and click OK.

An information dialog box is displayed.
9 Click OK.
The cluster is moved.
Remove a cluster from a cluster folder.
10 In the navigation tree, right-click the cluster and choose Remove.
An information dialog box is displayed.
11 Click OK.
An information dialog box is displayed.
12 Click OK.
The cluster is removed.
Remove a cluster folder.
13 In the navigation tree, right-click the cluster folder and choose Remove.
An information dialog box is displayed.
14 Click OK.
An information dialog box is displayed.
15 Click OK.
The cluster folder is removed.
----End

2.5 Configuring the Computing Resource Scheduling

Policies
Scenarios
On FusionCompute, configure the policies for scheduling computing resources in a cluster to
implement dynamic resource scheduling and load balancing.
Dynamic Resource Scheduling
The dynamic resource scheduling function enables the system to use an intelligent load
balancing scheduling algorithm to periodically check the load on each physical host in a
cluster and migrate VMs among hosts as needed. In this manner, the system can balance load
among hosts in the cluster. This function has the following characteristics:
l The migration automation level of a cluster can be set to automatic or manual. In
automatic mode, the system automatically migrates VMs to the most appropriate host. In
manual mode, the system provides an operation suggestion for the system administrator
to determine whether to perform the operation as suggested.
l Resources can be scheduled by CPU, memory, or CPU plus memory. The VM
performance is affected only when the memory overcommitment function is enabled.
Therefore, if memory overcommitment is not enabled, scheduling by CPU is
recommended to balance CPU usage in the cluster. If memory overcommitment is
enabled, scheduling by CPU plus memory is recommended.
l Advanced scheduling rules can be used to meet special resource requirements. For
example, a VM-mutually-exclusive rule can be configured for two VMs that work in

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 14

active/standby mode to ensure that the VMs run on different hosts. This improves system
reliability.
l The time-based scheduling settings allow the system to schedule resource in different
time periods based on service requirements. VM migration requires certain system
overhead. Therefore, set a conservative policy for a heavy-traffic scenario and a medium
or radical policy for a light-traffic scenario to ensure the optimal service performance.
If resources in a cluster cannot be scheduled, rectify this problem by following operations
provided in Failed to Execute the Policy for Scheduling Computing Resources in a
Cluster.

NOTE
To maximize resource scheduling efficiency, make sure that hosts in the same cluster use the same CPU,
memory, network, and storage configurations, so that VMs can be migrated across the hosts in the
cluster when required.

VM Scheduling Customization
The VM scheduling customization function allows users to set a scheduling automation level
for each VM in a dynamic resource scheduling (DRS) cluster to meet diverse VM scheduling
automation level requirements.
The automation level customized for a VM prevails over the automation level set for the
cluster of the VM. You can set different automation levels for a VM and its cluster. If the
automation level of a VM is set to Disabled, the system will not migrate this VM or provide
any migration suggestions for it. If the VM scheduling customization function is disabled or
the automation level of a VM is set to Default, the VM uses the same automation level as the
cluster.
The system will not migrate VMs that do not meet migration requirements. For details about
the VM migration requirements, see Appendix > FAQ > VM Migration Requirements in
the FusionCompute V100R005C10 Virtual Machine Management Guide.
Automatic Power Management
The automatic power management function enables the system to periodically check resource
usage on hosts in a cluster. If resources in the cluster are sufficient but service load on each
host is light, the system migrates VMs on redundant hosts and powers off these hosts to
reduce power consumption. If the in-service hosts are overloaded, the system powers on
offline hosts in the cluster to balance load among hosts.
l When the automatic power management function is enabled, automatic computing
resource scheduling must be enabled so that the system automatically balances VM load
on hosts after the hosts are powered on.
l The time-based power management settings allow the system to manage power in
different time periods based on service requirements. When services are running stably,
set automatic power management to a low level to prevent adverse impact on services.
l With the automatic power management function enabled, the system checks resource
usage in the cluster, and powers off some light loaded hosts only when the resource
utilization drops below the light-load threshold over the specified time period (the
default value is 40 minutes). Similarly, the system powers on some hosts only when the
resource utilization rises above the heavy-load threshold over the specified time period
(the default value is 5 minutes). You can customize the time period for evaluating the
thresholds of powering on or off hosts.
For example, the system administrator can enable automatic power management for a holiday
or off-duty period that requires less resources than usual. The system will power off some

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 15

hosts in this period to reduce energy consumption. When users start to use VMs on working
days, the system will automatically power on hosts to meet increasing resource requirements.

NOTE
A several-minute delay is required before a host is powered on. If you perform an operation that requires
a large amount of resources during this delay period, for example, start VMs, the operation may fail due
to resource insufficiency. In this case, perform the operation again after several minutes.

Resource Scheduling Interval

The default resource scheduling interval for both computing resource scheduling and
automatic power management is 10 minutes, which is configurable to administrators. For
details, see Configuring the Resource Scheduling Interval.

Balancing Group

Users can add VMs in the current cluster to the balancing group. After resources are
scheduled, resources in the cluster are balanced among hosts. The balancing group enables
VMs in the group are evenly distributed in the cluster.

Advanced Settings

FusionCompute provides a computing resource scheduling algorithm, which supports

advanced settings to meet requirements of different scenarios.

Before you set any options related to advances settings, contact Huawei technical support
engineers and have them evaluate and confirm the settings to ensure the resource scheduling
efficiency and effects.

Prerequisites
Conditions
l Computing, storage, and network resources on hosts in a cluster are shared.
l You have logged in to FusionCompute.
l You have obtained the name of the cluster for which the resource scheduling rules are to
be configured.
l You have configured the baseboard management controller (BMC) for each host before
configuring automated power management. For details, see Configuring Host BMC
Parameters.

Procedure
Configure the basic information.
1 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
2 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.
3 Click Configuration and click Control Cluster Resource on the right pane of the page.
A configuration page is displayed.
4 Click Configure Computing Resource Scheduling.
The Configure Computing Resource Scheduling page is displayed.
5 Select Enable computing resource scheduling.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 16

After this parameter is selected, parameters on the Configure Resource Scheduling

page are available to configuration. Resource Group, Rule Group, Balancing Group,
and Advanced Settings are displayed in the navigation tree, as shown in Figure 2-1.
To disable computing resource scheduling, deselect Enable computing resource
scheduling and click OK to make the settings take effect.

Figure 2-1 Configure Resource Scheduling

6 Set the following parameters on the Configure Resource Scheduling page:

Automated Computing Resource Scheduling
– Automation level: Set it to Manual or Automatic. Manual indicates that VM
migration recommendations are provided for the administrator when the migration
threshold is reached. Automatic indicates that the system automatically migrates
VMs based on set thresholds.
– Measured by: Set it to CPU, Memory, or CPU & Memory. If this parameter is set
to CPU, VM migration is performed based on CPU usage. If this parameter is set to
Memory, VM migration is performed based on memory usage. If this parameter is
set to CPU & Memory, VM migration is performed based on CPU and memory
usage.
– Migration: Specifies the baseline for VM migration on hosts in the cluster. If the
load on a host is lower than the set baseline, VMs on the host will not migrate to
another host due to the load balance mechanism.
The migration baseline configuration may affect the automated power management
function. Therefore, do not set the migration baseline when the power management
function is enabled.
– Migration Threshold: Each column in the table indicates a one-hour period. You
can set a migration threshold for each one-hour period.
– Resource scheduling interval: Set the period during which the specified
scheduling thresholds take effect.
– VM scheduling customization: If this parameter is selected, you can set an
automation level for each VM in the cluster. If this parameter is not selected, the
VMs in a cluster use the automation level specified for the cluster.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 17

Power Management Parameters

– Enable automated power management: If this parameter is selected, the system
migrates the VMs on a host and powers on or off some hosts in the cluster based on
the host resource usages. This function is remotely implemented by the host BMC.
Therefore, you need to configure the BMC for all hosts in the cluster.
Automated power management depends on automated computing resource
scheduling. Therefore, automated power management takes effects only when
automated computing resource scheduling is enabled and the migration threshold is
not set to Conservative or Radical.
– Automation Levels
n Manual: The user is prompted to migrate VMs based on the suggestions
provided on the Computing Resource Scheduling page.
n Automatic: The system automatically migrates VMs to maximize resource
utilization when the resource load is heavy.
– Power Management Thresholds
n Conservative: The system does not power off hosts by default. It powers on
available hosts in the cluster only when the average host resource usage in the
cluster is higher than the heavy-load threshold.
n Slightly conservative, Medium and Slightly radical: The system powers on
available hosts in the cluster when the average host resource usage in the
cluster is higher than the heavy-load threshold. It powers off some hosts when
the average host resource usage in the cluster is lower than the light-load
threshold.
n Radical: The system does not power on available hosts by default. It powers
off some hosts in the cluster only when the average host resource usage in the
cluster is lower than the light-load threshold.
Table 2-2 lists the light-load and heavy-load threshold value of each power
management threshold.

Table 2-2 Light-load and heavy-load threshold value of each power management
threshold
Threshold Name Heavy-Load Light-Load Threshold
Threshold Value Value

Conservative 63% N/A

Slightly conservative 72% 23%

Medium 81% 45%

Slightly radical 90% 54%

Radical N/A 63%

– Threshold Validity Period Setting

The threshold validity period can be set to:
n By day: The thresholds always take effect.
n By week: The thresholds take effect only on the specified days in a week.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 18

n By month: The thresholds take effect only on the specified days in a month.
Configure a resource group.
To allow VMs in a specified VM group to be migrated only within a specific host group, the
VM group and the host group must be associated and the VM group members must be able to
run on the specified hosts in the host group. Therefore, configure VM group and host group
settings on the Resource Group page in advance based on the data plan.
7 Check whether the required VM group is displayed on the Resource Group page.
If yes, go to 14.
If no, go to 8.
8 On the Resource Group page, click Add in the VM Group area.
After adding VMs in a cluster to a VM group, you can set a VM migration range for the
VM group by configuring a VMs to hosts rule.
The Add VM Group page is displayed, as shown in Figure 2-2.

Figure 2-2 Add VM Group

9 Enter the VM group name.

10 Enter the search criteria in Available VMs, and click Search.
The VM group that meets the search criteria is displayed.
Search criteria can be its name and ID.

11 Select the VMs to be added to the VM group and click to add them to the list
in the right pane.
12 Click OK.
A dialog box is displayed.
13 Click OK.
The VM group is added.
To modify or delete the VM group, locate the row that contains the VM group on the
Resource Group page, click Modify or Delete.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 19

14 Check whether the required host group is displayed on the Resource Group page.
If yes, go to 22.
If no, go to 15.
15 On the Resource Group page, click Add in the Host Group area.
After adding hosts in a cluster to a host group, you can set a VM migration range for
VMs by configuring a VMs to hosts rule.
A dialog box shown in Figure 2-3 is displayed.

Figure 2-3 Add Host Group

16 Enter the host group name.

17 Enter the search criteria in Available Hosts, and click Search.
The host group that meets the search criteria is displayed.
Search criteria can be its name and IP address.

18 Select the hosts to be added to the host group and click to add them to the list
in the right pane.
19 Click OK.
A dialog box is displayed.
20 Click OK.
The host group is added.
To modify or delete the host group, locate the row that contains the host group on the
Resource Group page, click Modify or Delete.
Configure scheduling rules.
21 Perform the corresponding operation based on the rule specified.
– To keep the selected VMs running on the same host, go to 32.
– To keep the selected VMs running on different hosts, go to 27.
– To limit the migration range of VMs in a specified VM group to a specific host
group, go to 22.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 20

Keep VMs together: This rule keeps the selected VMs always running on the same host.
One VM can be added to only one keep-VMs-together rule.
Mutually exclusive VMs: This rule keeps the selected VMs running on different hosts.
One VM can be added to only one VM-mutually-exclusive rule.
VMs to hosts: This rule associates a VM group with a host group so that the specified
VMs in the VM group can be configured to run only on the specified host in the host
group.
22 Click Add on the Rule Group page.
A dialog box shown in Figure 2-4 is displayed.

Figure 2-4 Add Rules

23 Enter a rule name in Name and set Type to VMs to Hosts.

VMs to hosts: This rule associates a VM group with a host group so that the specified
VMs in the VM group can be configured to run only on the specified host in the host
group.
24 Select the VM group and host group to be associated with each other and set Rule to
either of the following values based on the service requirements:
– Must run on hosts in group: VMs in the VM group must run on the hosts in the
host group.
– Should run on host group: VMs in the VM group are advised to run on the hosts
in the host group.
– Must not run on hosts in group: VMs in the VM group must not run on the hosts
in the host group.
– Should not run on hosts in group: VMs in the VM group are advised not to run on
the hosts in the host group.
25 Click OK.
A dialog box is displayed.
26 Click OK.
The VMs to hosts rule is added.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 21

To modify or delete the rule, locate the row that contains the rule on the Rule
Management page, click Modify or Delete.
Go to 37 after this step.
27 Click Add on the Rule Management page.
A dialog box is displayed.
28 Enter a rule name in Name and set Type to Mutually exclusive VMs.
Mutually exclusive VMs: This rule keeps the selected VMs running on different hosts.
One VM can be added to only one VM-mutually-exclusive rule.

29 In the Available VMs area, select two VMs and click to move them to
Selected VMs.
30 Click OK.
A dialog box is displayed.
31 Click OK.
The Mutually exclusive VMs rule is added.
To modify or delete the rule, locate the row that contains the rule on the Rule
Management page, click Modify or Delete.
After this step, go to 37.
32 Click Add on the Rule Management page.
A dialog box is displayed.
33 Enter a rule name in Name and set Type to Keep VMs together.
Keep VMs together: This rule keeps the selected VMs always running on the same host.
One VM can be added to only one keep-VMs-together rule.

34 In the Available VMs area, select two VMs and click to move them to
Selected VMs.
35 Click OK.
A dialog box is displayed.
36 Click OK.
The Keep VMs together rule is added.
To modify or delete the rule, locate the row that contains the rule on the Rule
Management page, click Modify or Delete.
Configure balancing groups.

Users can add some VMs in the current cluster to a balancing group. The workload on these
VMs can be further balanced on the basis that the cluster workload is balanced. For example,
if two services are deployed on VMs in the same cluster, the VMs can be added to a balancing
group to ensure that the workload of these two services can be balanced among these VMs.
37 On the Balancing Group page, click Add.
A dialog box is displayed.
38 Set a name for the group.

39 Select VMs from the Available VMs area and click to move the selected VMs
to the Selected VMs area.
40 Click OK.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 22

A dialog box is displayed.

41 Click OK.
The balancing group is added.
Configure advanced settings.
42 On the Advanced Settings page, click Add.
A dialog box is displayed.
43 Enter an advanced setting name and its value.
For details about the algorithm parameters, see Parameters in Advanced Settings for
Computing Resource Scheduling. For details about setting rules, contact technical
support.
44 Click OK.
The advanced settings are configured.
----End

2.6 Configuring the HA Policy for a Cluster

Scenarios
On FusionCompute, configure the high availability (HA) policy for a cluster to ensure that
sufficient resources are available in the cluster to fail over VMs on a faulty host if any.

Prerequisites
Conditions
You have logged in to FusionCompute.

Procedure
Step 1 On FusionCompute, click Computing Pool.

Step 2 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.
Step 3 Choose Configuration > Configure HA and click Control Cluster Resource.
A dialog box is displayed, as shown in Figure 2-5.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 23

Figure 2-5 Set Cluster Resource Control

Step 4 Enable is selected by default.

The HA function can be enabled for VMs in a cluster only when this function is enabled for
the cluster.

Step 5 Select HA resource reservation, Use dedicated failover hosts, or Tolerate cluster host
failures.
l HA resource reservation: specifies whether the system reserves preset CPU and
memory resources in the cluster to implement VM HA.
l Use dedicated failover hosts: specifies whether system reserves some hosts as dedicated
failover hosts. VMs that are working properly cannot be started, woken up, or restored
using a snapshot on a failover host, or migrated to it.
If an HA-enabled VM fails, the system restarts the VM on a common host or a failover
host based on resource usage on the hosts.
If Migrate all VMs on the dedicated failover host is selected, the system periodically
migrates VMs on a failover host to other common hosts that have sufficient resources to
reserve resources on the failover host.
l Tolerate cluster host failures: specifies the number of tolerated faulty hosts in a cluster.
The system periodically checks whether resources in the cluster are sufficient for VMs
failover on these hosts. If the resources are insufficient, the system generates an alarm to
notify the user.
A slot is a logical unit of memory and CPU resources. You can set slot sizes in
Automatic or Custom mode.
– Automatic: The system selects the maximum reserved CPU resources (MHz) and
memory resources (MB/GB) on all running VMs in the cluster as the CPU slot size
and the memory slot size.
– Custom: You need to set a custom CPU slot size and a memory slot size based on
site requirements.

Step 6 Perform the required operation based on the selected HA policy.

l If HA resource reservation is selected, go to Step 7.
l If Use dedicated failover hosts is selected, go to Step 8.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 24

l If Tolerate cluster host failures is selected, go to Step 11.

Step 7 Set Reserved CPU (%) and Reserved memory (%).
l Reserved CPU (%): specifies the percentage of the CPU resources reserved to the total
CPU resources in the cluster.
l Reserved memory (%): specifies the percentage of the memory resources reserved to
the total memory resources in the cluster.
After this step, go to Step 12.
Step 8 Click Add.
A dialog box is displayed.
Step 9 Select a host as a failover host and click OK.

Step 10 To allow all VMs to be migrated from the failover host periodically, select Migrate all VMs
on the dedicated failover host.
After this step, go to Step 12.
Step 11 Set the number of tolerated faulty hosts and the slot size based on the following configuration
principles:
1. Calculate the slot sizes.
A slot is a logical unit of memory and CPU resources.
– If Automatic is selected, the system selects the maximum CPU resources (MHz)
and memory resources (MB) on all running VMs in the cluster as the CPU slot size
and the memory slot size.
– If Custom is selected, you need to set a custom CPU slot size and a memory slot
size based on site requirements. This mode is recommended when the cluster
contains any VMs that have much larger reservations than the others.
To set proper slot sizes, perform the following operations:
Enter the maximum CPU resources (MHz) and memory resources (MB) on all
running VMs in the cluster as the CPU slot size and the memory slot size, click
Calculate, and view how many VMs require multiple slots. Modify the slot sizes
based on the calculated value.
2. Calculate the total capacity of the cluster.
The total capacity indicates the maximum number of slots that can be supported in the
cluster.
Capacity of a host = Rounddown [Min (Host CPU capacity/CPU slot size, Host memory
capacity/Memory slot size)]
The total capacity of a cluster is the sum of the capacity of all hosts in the cluster.
3. Calculate the reserved capacity of the cluster for VM failover.
Reserved capacity of a cluster for VM failover = Total capacity of the cluster – Total
capacity of first N hosts that have largest reservations (N is the number of tolerated
faulty hosts)
The reserved capacity of a cluster for VM failover must be greater than the number of
running VMs in the cluster. You can adjust the number of tolerated faulty hosts based on
this formula.
For example, a cluster has three hosts, H1 (9 GHz CPU, 9 GB memory), H2 (9 GHz CPU, 6
GB memory), and H3 (6 GHz CPU, 6 GB memory), and five running VMs, VM1 (2 GHz

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 25

CPU, 1 GB memory), VM2 (2 GHz CPU, 1 GB memory), VM3 (1 GHz CPU, 2 GB

memory), VM4 (1 GHz CPU, 1 GB memory), and VM5 (1 GHz CPU, 1 GB memory).

In this case, the maximum CPU resources and memory resources are 2 GHz and 3 GB. The
capacity of H1 is 4 (Rounddown {Min [9/2, 9/2]}); the capacity of H2 is 3 (Rounddown {Min
[9/2, 6/2]}); the capacity of H3 is 3 (Rounddown {Min [6/2, 6/2]}). Therefore, the total
capacity of the cluster is 10 (4 + 3+3). If the number of tolerated faulty hosts is set to 1, the
reserved capacity of a cluster for VM failover is 6 (10 – 4). 6 is greater than the number of
running VMs (5). If the number of tolerated faulty hosts is set to 2, the reserved capacity of a
cluster for VM failover is 3 (10 – 4 – 3). 3 is less than the number of running VMs (5).
Therefore, 1 is recommended for the number of tolerated faulty hosts.

Step 12 Click OK.

An information dialog box is displayed.

Step 13 Click OK.

The HA resource reservation function is enabled.

----End

2.7 Configuring the VM Processing Policy upon Data

Store Faults

Scenarios
On FusionCompute, configure the VM processing policy upon data store faults for a cluster.
The policy can be set to:

If the data store type is FusionStorage, the cluster does not support the data store
troubleshooting policy configuration.

l (Default) No processing: If a data store becomes faulty, the system does not handle VMs
that use the data store. The VMs will be automatically started after the fault is rectified.
l Stop VM: If a data store becomes faulty, the system stops all VMs that use the data
store. This policy does not apply to data stores of the FusionStorage type.

Prerequisites
Conditions
You have logged in to FusionCompute.

Procedure
Step 1 On FusionCompute, click Computing Pool.

Step 2 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.

Step 3 Select Configuration > Basic Configuration > Control Cluster Resource.
A dialog box is displayed, as shown in Figure 2-6.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 26

Figure 2-6 Advanced Settings

Step 4 Set VM processing policy upon data store faults: to No processing or Stop VM.
l (Default) No processing: If a data store becomes faulty, the system does not handle VMs
that use the data store. The VMs will be automatically started after the fault is rectified.
l Stop VM: If a data store becomes faulty, the system stops all VMs that use the data
store. This policy does not apply to data stores of the FusionStorage type.

Step 5 Click OK.

An information dialog box is displayed.

Step 6 Click OK.

The policy is configured.

----End

2.8 Enabling Memory Overcommitment in a Cluster

Scenarios
On FusionCompute, enable the host memory overcommitment function in the cluster.
Memory overcommitment allows more space than the physical host has available to be used
by VMs, improving VM density on a host. The enabled Guest NUMA function become
unavailable if host memory overcommitment is enabled or the host CPU resource mode is set
to isolated.

The memory overcommitment function cannot be enabled for a cluster if the cluster contains
hosts that use intelligent network cards (iNICs).

Impact on the System

The VM performance deteriorates on a host that has the memory overcommitment function
enabled.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 27

Prerequisites
Conditions
You have logged in to FusionCompute.

Procedure
Step 1 On FusionCompute, click Computing Pool.

Step 2 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.
Step 3 Click the Configuration tab and then click Control Cluster Resource.
A dialog box shown in Figure 2-7 is displayed.

Figure 2-7 Set Cluster Resource Control

Step 4 Set Host memory overcommitment to Enable.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 28

NOTE

l After host memory overcommitment is enabled, the overcommitment ratio for a VM can be
controlled by specifying the Reserved (MB) parameter in the VM QoS settings area.
– If the configured VM memory is greater than or equal to 16 GB, the VM Reserved (MB)
parameter must be set to its maximum value to ensure the optimal VM performance. In this
case, host memory overcommitment does not take effect for the VM.
– If the configured VM memory is less than 16 GB, the VM Reserved (MB) value can be set to
70% of the specified VM memory size so that more VMs than the number of physically
supported ones can use the host memory resources in practice. If the monitored VM memory
usages have been persistently greater than 40% in several hours, the VM Reserved (MB)
parameter must be set to its maximum value. In this case, host memory overcommitment does
not take effect for the VM.
l After host memory overcommitment is enabled, the total available memory capacity equals to the
total memory capacity in the virtualization domain multiplied by 120%. The total memory capacity
in the virtualization domain equals to the server memory capacity minus memory size required by
virtualization management. You can choose Computing Pool > Site > Cluster > Host > Summary
> Monitoring Information to view the total memory capacity in the virtualization domain.
l After host memory overcommitment is enabled, plan VMs on hosts based on the total memory
capacity. If VMs that consumed a large amount of memory exist, some VMs may fail to start even if
the used memory has been released. This startup failure occurs because the virtualization layer does
not know that the memory is released.
l After host memory overcommitment is enabled, VMs cannot be hibernated and memory snapshots
cannot be created for them.

Step 5 Click OK.

A dialog box is displayed.
Step 6 Click OK.
The host memory overcommitment function is enabled.

----End

2.9 Configuring the VM Startup Policy

Scenarios
On FusionCompute, configure the VM startup policy, which allows the system to select a
proper host in a cluster to start a VM.
If computing resource scheduling is enabled for the cluster, the system uses the default VM
startup policy Assign based on load balancing.

Prerequisites
Conditions
You have logged in to FusionCompute.

Procedure
Step 1 On FusionCompute, click Computing Pool.

Step 2 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 29

Step 3 Click the Configuration tab and then click Control Cluster Resource.
A dialog box shown in Figure 2-8 is displayed.

Figure 2-8 Set Cluster Resource Control

Step 4 Configure the VM startup policy.

l Assign automatically: The system starts a VM on a host that has available resources in
the cluster.
l Assign based on load balancing: The system starts a VM on a host that has lower CPU
and memory usage.

Step 5 Click OK.

A dialog box is displayed.

Step 6 Click OK.

The VM startup policies are configured.

----End

2.10 Configuring Guest NUMA for a Cluster

Scenarios
On FusionCompute, configure the Guest Non-Uniform Memory Access (NUMA) function for
hosts in a cluster. Guest NUMA presents a topology view of memory and CPU resources on
each host to VMs, the VM user can configure VM CPUs and memory using the third-party
software (such as Eclipse) based on this topology, so that VMs can obtain the most easy-to-
access memory based on the topology, thereby reducing access latency and improving VM
performance.

The enabled Guest NUMA function become unavailable if host memory overcommitment is
enabled or the host CPU resource mode is set to isolated.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 30

NOTE
Ensure that the following conditions are met for the Guest NUMA function to take effect:
l The number of VM CPUs supported by a host in the cluster must be a multiple of the physical CPU
quantity on the host or a multiple of the thread quantity of a single CPU of the host. (To check the
number of physical CPUs on a host or the number of CPU threads of a host, switch to the host
hardware information page and choose Hardware > CPU, as described in Querying Host
Information.)
The Guest NUMA function may fail on the host that accommodates the VM if the host CPU
quantity changes (due to VM live migration or VM startup on another host) or VM CPU
specifications change.
l The memory overcommitment function or the host CPU resource mode is disabled for the cluster.
l The NUMA function is enabled for hosts in the cluster. For example, to enable the NUMA function
for a RH2288H V2 server, choose Advanced > Advanced Processor in the advanced basic input/
output system (BIOS) settings of the server, and set NUMA mode to Enabled.
l VMs on the hosts in the cluster are restarted after the Guest NUMA function is enabled for the
cluster.

Prerequisites
Conditions
l You have logged in to FusionCompute.
l You have enabled the Host NUMA function in the VM BIOS.

Procedure
Step 1 On FusionCompute, click Computing Pool.

Step 2 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.
Step 3 Click Advanced Settings.
A dialog box shown in Figure 2-9 is displayed.

Figure 2-9 Advanced Settings

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 31

Step 4 Deselect Host memory overcommitment.

Step 5 Select GuestNUMA.
Step 6 Click OK.
A dialog box is displayed.
Step 7 Click OK.
The Guest NUMA function is configured for the cluster.
NOTE
Guest NUMA cannot be enabled or disabled for the cluster when the cluster contains VMs in a transition
state, such as running or migrating.

Step 8 Restart VMs in the cluster to make the configurations take effect.
For details, see Restarting a VM in VM Operation Management in the FusionCompute
V100R005C10 Virtual Machine Management Guide.
----End

2.11 Configuring IMC for a Cluster

Scenarios
On FusionCompute, configure the incompatible migration cluster (IMC) function for a cluster
to enable VM migration across hosts that use CPUs of different performance baselines in the
cluster.
Currently, the IMC function applies to the CPUs from Intel only.
The IMC function allows the hosts in a cluster to present the same CPU function set to VMs
running on them. This function ensures successful VM migration across these hosts even
when the hosts physically use CPUs of different performance baselines.
Ensure that the following conditions are met for the IMC function if the cluster contains hosts
and VMs:
l The CPU generations of the hosts in the cluster are the same as or later than the target
IMC mode.
l The CPU generations of the running or hibernated VMs in the cluster are the same as or
earlier than the target IMC mode. If any VM in the cluster does not meet this
requirement, the VM must be stopped or migrated to another cluster.

NOTE
After setting an IMC mode for a cluster, you need to enable the Execute Disable Bit function, which is
also known as the No eXecute bit (NX) or eXecute Disable (ND) function, in the BIOS advanced CPU
options for existing hosts in the cluster and the hosts to be added to the cluster.

Prerequisites
Conditions
You have logged in to FusionCompute.

Procedure
Check the IMC mode of each host in the cluster.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 32

1 On FusionCompute, click Computing Pool.

2 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.

3 On the Host page, click on the right part of the host list, and select Select Column.
A dialog box is displayed.
4 Select IMC and Latest Supported IMC CPU Generation and click OK.
The host list displays the IMC and Latest Supported IMC CPU Generation columns.
IMC specifies the IMC mode current used by a host, and Latest Supported IMC CPU
Generation specifies the latest IMC CPU generation supported by a host.
5 Identify the lowest IMC mode in the Latest Supported IMC CPU Generation column.
FusionCompute supports the following Intel CPU generations, which present higher
performance levels in ascending order:
– Merom
– Penryn
– Nehalem
– Westmere
– Sandy Bridge
– Ivy Bridge
Set the IMC mode for the cluster.
6 Click Advanced Settings on the Getting Started tab.
A dialog box shown in Figure 2-10 is displayed.

Figure 2-10 Advanced Settings

7 Click Enable IMC and set IMC Mode to the lowest IMC mode identified in the Latest
Supported IMC CPU Generation column.
The IMC mode of a cluster must be set to the CPU generation that exposes the minimum
function set in the cluster or an earlier version of the CPU generation.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 33

8 Click OK.
A dialog box is displayed.
9 Click OK.
The IMC mode is configured for the cluster.
After the configuration, you can check the IMC settings of the cluster in the IMC
column of the host on the Host page.
Follow-up Procedure
After setting an IMC mode for a cluster, you need to enable the Execute Disable Bit function,
which is also known as the No eXecute bit (NX) or eXecute Disable (ND) function, in the
BIOS advanced CPU options for existing hosts in the cluster and the hosts to be added to the
cluster.
After setting an IMC mode for a cluster, you can migrate VMs across hosts that use CPUs of
different performance baselines. For details, see Migrating a VM in VM Operation
Management in the FusionCompute V100R005C10 Virtual Machine Management Guide.

----End

2.12 Removing a Cluster

Scenarios
On FusionCompute, remove an unwanted user cluster.

Prerequisites
Conditions
l You have logged in to FusionCompute.
l The hosts in the cluster have been removed.
Data
The name of the cluster to be removed is available.

Procedure
Step 1 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.

Step 2 In the navigation tree on the left, select the site and cluster.

Step 3 Right-click the cluster to be removed, and choose Remove.

A dialog box is displayed.

Step 4 Click OK.

A dialog box is displayed.

Step 5 Click OK.

The cluster is removed.

----End

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 34

3 Host Management

About This Chapter

3.1 Adding Hosts

3.2 Querying Host Information
3.3 Modifying the Name and Description of a Host
3.4 Modifying the Host name in OS
3.5 Changing the Switching Mode of a Network Port
3.6 Updating the Number of Virtual Ports of the SR-IOV-enabled Network Port
3.7 Setting a Host to the Maintenance Mode
3.8 Changing the Multipathing Type of a Host
3.9 Setting Time Synchronization on a Host
3.10 Performing Forcible Time Synchronization
3.11 Configuring Host BMC Parameters
3.12 Sharing the GPU on a Host
3.13 Configuring the Host GPU Mode
3.14 Enabling or Disabling LLDP for a Host
3.15 Enabling the Host Antivirus Function
3.16 Moving a Host
3.17 Removing a Host

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 35

3.1 Adding Hosts

Scenarios
On FusionCompute, add hosts to a cluster.

When you add hosts to a cluster, the system automatically binds the management ports in
active/standby mode. Therefore, you need to delete the Eth-Trunk configuration on the access
switch. Otherwise, network communication is interrupted.

After hosts are added to a cluster, add other planned management network ports to the bound
network port to improve reliability of the management plane network.

Prerequisites
Conditions
l You have logged in to the FusionCompute.
l The required operating system (OS) is installed on the server carrying the host. For
details, see the FusionCompute V100R005C10 Software Installation Guide.
l The Virtualization Resource Management (VRM) management plane has been initialized
after FusionCompute installation is complete, and the communication between the host
and the Virtualization Resource Management (VRM) management plane is normal. Only
one management plane can be configured for the VRM node.
l The CPU generation of the host is the same as or later than the incompatible migration
cluster (IMC) mode configured for the cluster if the target cluster has IMC mode
enabled. Execute Disable Bit function, which is also known as the No eXecute bit (NX)
or eXecute Disable (ND) function, has been enabled in the advanced CPU options for the
BIOS of the host.

Procedure
Determine the method of adding hosts.

1 Determine the method of adding hosts.

– To add hosts in batches using a template, go to 7.
All information about the hosts can be set in a template and imported at a time. This
method is recommended when a large number of hosts need to be added.
– To manually add hosts one by one, go to 2.
Only one host can be added at a time. This method is easy to operate and is
recommended when a small number of hosts are to be added.
The Add Host page is displayed.

2 On FusionCompute, click Computing Pool.

The Computing Pool page is displayed.
3 In the navigation tree on the left, select the site, right-click the cluster, and choose Add
Host.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 36

NOTICE
The Connect to OpenStack option can be configured only when no host is added. If
hosts have been added to the FusionCompute system, you are not allowed to change the
Connect to OpenStack option.

A dialog box is displayed, as shown in Figure 3-1.

Figure 3-1 Add Host

Add a host.
4 Set the following parameters for the host:
– Name
– IP address
– BMC IP
– Username
– Password
– Use the site time sync policy: synchronizes the configured cluster system time to
the host.
5 Click OK.
An information dialog box is displayed.
6 Click OK.
The host is added to the cluster.
After this step is complete, no further action is required.
Obtain the template for adding hosts.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 37

7 On FusionCompute, click Computing Pool.

The Computing Pool page is displayed.
8 In the navigation tree, select the site and cluster.
The Getting Started page of the cluster is displayed.
9 In the upper part of the page, click Operation and select Add Hosts in Batches.
The Add Hosts in Batches page is displayed, as shown in Figure 3-2.

Figure 3-2 Add Hosts in Batches

10 Click Download Template.

The page for downloading templates is displayed.
11 Download the template to the PC.
The default file name of the template is HostYYYY-MM-DD HHMMSSXXX.xls.
12 Open the template and set the following parameters for the host:
– Host IP Address: Enter the management IP address of the host.
– Host Name
– Host Description
– BMC IP Address: Enter the BMC IP address of the host. This parameter is
required if power management automation needs to be set.
– BMC Username: Enter the username for logging in to the BMC of the host. This
parameter is required if power management automation needs to be set.
– BMC Password: Enter the password for logging in to the BMC of the host. This
parameter is required if power management automation needs to be set.
– Use the site time sync policy: If you set this parameter to Yes, the site time
synchronization policy is automatically applied on the host after the host is
successfully added. This automatic policy synchronization requires the restart of
host management service processes, which will interrupt host services for 3 to 5
minutes. If you set this parameter to No, you can manually configure a time
synchronization policy for the host after it is added.
13 Save and close the template.
14 Click Browse to the right of Import template file on the Add Hosts in Batches page.
A dialog box is displayed.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 38

15 Select the template and click Open.

16 Click OK.
An information dialog box is displayed.
17 Click OK.
You can choose System > Tasks and Logs > Task Center to view the task progress.
Follow-up Procedure
To configure an external clock source, configure the time synchronization function on the
host. For details, see Setting Time Synchronization on a Host and Performing Forcible
Time Synchronization.
Configure storage for the hosts. To use Huawei storage area network (SAN) storage devices,
modify the storage multipathing type of the hosts and then add storage ports to the hosts. To
use other types of storage devices, simply add storage ports to the hosts. For details, see
Changing the Multipathing Type of a Host and Adding a Storage Port.
Configure networks for the hosts. If a distributed virtual switch (DVS) is not available, create
a DVS and then add uplinks to the hosts. If a DVS is available, simply add uplinks to the
hosts. For details, see Distributed Virtual Switch Management > Creating a Distributed
Virtual Switch or Upstream Link Group Management > Adding an Upstream Link in the
FusionCompute V100R005C10 Network Management Guide.
----End

3.2 Querying Host Information

Scenarios
On FusionCompute, query the information about a host.

Prerequisites
Conditions
You have logged in to FusionCompute.
Data
You have obtained the name of the host to be queried.

Procedure

1 Perform the following operations based on the information to be queried:

– To query the general information about the host, go to 2.
– To query the information about VMs on the host, go to 5.
– To query the information about devices used by the host, go to 8.
– To query the information about data stores associated with the host, go to 11.
– To query the information about hardware used by the host, go to 14.
Query the general information about the host.
2 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 39

3 In the navigation tree on the left, select the site, the cluster, and the host.
The Getting Started page is displayed.
4 Query the general information about the host on the Summary page.
No further action is required.
Query the information about VMs on the host.
5 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
6 In the navigation tree on the left, select the site, the cluster, and the host.
The Getting Started page is displayed.
7 Query the information about VMs on the host on the VM page.
No further action is required.
Query the information about devices used by the host.
8 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
9 In the navigation tree on the left, select the site, the cluster, and the host.
The Getting Started page is displayed.
10 Query the information about devices used by the host on the Device page.
No further action is required.
Query the information about data stores associated with the host.
11 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
12 In the navigation tree on the left, select the site, the cluster, and the host.
The Getting Started page is displayed.
13 Query the information about data stores associated with the host on the Storage page.
No further action is required.
Query the information about hardware used by the host.
14 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
15 In the navigation tree on the left, select the site, the cluster, and the host.
The Getting Started page is displayed.
16 Query the information about hardware used by the host on the Hardware page.
No further action is required.

----End

3.3 Modifying the Name and Description of a Host

Scenarios
On FusionCompute, modify the name and description of a host.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 40

Prerequisites
Conditions
You have logged in to FusionCompute.

5 In the Basic Information area on the Summary page, click in the Name and
Description rows.
A dialog box is displayed.
6 Enter the new name and description of the host.
7 Click OK.
The host information is modified.
----End

3.4 Modifying the Host name in OS

Scenarios
On FusionCompute, modify the Host name in OS.
NOTE
Historical logs of a host cannot be identified or backed up after the host name recorded in the OS is
changed. Therefore, back up the logs before changing the name or delete them if they are no longer
used.

Prerequisites
Conditions
You have logged in to FusionCompute.

Procedure
Search for the target host.
1 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 41

2 In the navigation tree on the left, select the target site.

The Getting Started page is displayed.
3 On the Host page, enter the search criteria and click Search.
The host that meets the search criteria is displayed.
Search criteria can be the host name.
Modify the host name in OS.
4 Click the name of the target host in the host list.

5 In the Basic Information area on the Summary page, click in the Host name in
OS row.
A dialog box is displayed.
6 Enter the new host name in OS.
7 Click OK.
A dialog box is displayed.
8 Click OK.
The host name in OS is modified.

----End

3.5 Changing the Switching Mode of a Network Port

Scenarios
On FusionCompute, change the switching mode of a network port between Standard and
SR-IOV-enabled when Intel 82599 network interface cards (NICs) are used. If the MZ510 or
MZ512 NICs are used, change the switching mode of a network port by following the steps
provided in Manually Changing the Switching Mode of a Network Port.

Prerequisites
Conditions
l The VT-x/VT-d support and the SR-IOV function have been enabled for the host
accommodating the network port for which the switching mode is to be changed. For
details about how to enable the support and the function, see Enabling the VT-x/VT-d
Support and SR-IOV Function.
l The network port for which the switching mode is to be changed is not in use.
l You have obtained the name of the network port for which the switching mode is to be
changed and the name of the host accommodating the network port.
l You have logged in to FusionCompute.

Procedure

1 On FusionCompute, click Computing Pool.

The Computing Pool page is displayed.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 42

2 In the navigation tree on the left, select the site, cluster, and host.
The Getting Started page is displayed.
3 Click Device.
4 Select PCI Device.
5 Click Change Switching Mode.
The Change Switching Mode dialog box is displayed.
6 Select the network port for which the switching mode is to be changed.
7 Click OK.
A dialog box is displayed.
8 Click OK.
The task is complete after the host restarts.
NOTE

If the Virtualization Resource Management (VRM) node is deployed on a VM in standalone mode

and the switching mode of a network port on the host accommodating the VRM node is changed,
you must run the following command to restart the NC process after the VRM VM starts:
service vnad restart

----End

Additional Information
Related Tasks
If a VM uses the distributed virtual switch (DVS) in SR-IOV-enabled mode, perform the
following steps to install the required NIC driver after the VM starts:
The following operations use the Windows operating system (OS) as an example:
1. Download the required NIC driver.
– If the Intel 82599 NIC is used, visit https://downloadcenter.intel.com/
Default.aspx?lang=zho to download the NIC driver.
– If the MZ510 or MZ512 NIC is used, visit http://www.emulex.com/downloads/
emulex.html to download the NIC driver.
2. On the Windows OS, double-click the NIC driver and install it as prompted.

3.6 Updating the Number of Virtual Ports of the SR-IOV-

enabled Network Port

Scenarios
On FusionCompute, update the number of virtual ports of a network port for which the
switching mode is set to SR-IOV-enabled when Intel 82599 network interface cards (NICs)
are used.

Prerequisites
Conditions
l The network port for which the number of virtual ports needs to be updated is not in use.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 43

l You have obtained the name of the network port for which the number of virtual ports
needs to be updated and the name of the host accommodating the network port.
l You have logged in to FusionCompute.

Procedure

1 On FusionCompute, click Computing Pool.

The Computing Pool page is displayed.
2 In the navigation tree on the left, select the site, cluster, and host.
The Getting Started page is displayed.
3 Click Device.
4 Select PCI Device.
5 Locate the row that contains the network port for which the number of virtual ports
needs to be updated and click Update Number of Ports.
The Update Number of Ports dialog box is displayed.
6 Set the port number.
The value ranges from 1 to 63.
7 Click OK.
A dialog box is displayed.
8 Click OK.
The task is complete after the host restarts.
NOTE

If the Virtualization Resource Management (VRM) node is deployed on a VM in standalone mode

and the switching mode of a network port on the host accommodating the VRM node is changed,
you must run the following command to restart the NC process after the VRM VM starts:
service vnad restart

----End

3.7 Setting a Host to the Maintenance Mode

Scenarios
On FusionCompute, set a host to maintenance mode. The host in maintenance mode is
isolated from the entire system, so that some maintenance operations, such as parts
replacement, power-off, or restart, can be performed on the host without affecting system
services.

NOTE
After a host enters maintenance mode, stop or migrate all VMs on the host, and then perform
maintenance operations.

Impact on the System

VMs cannot be created, started, or migrated on a host that works in maintenance mode.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 44

When a host works in the maintenance mode, the VMs on the host do not participate in
computing resource scheduling in the cluster that the VMs belong.

Prerequisites
Conditions
You have logged in to FusionCompute.

Procedure
Search for the target host.
1 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
2 In the navigation tree on the left, select the target site.
The Getting Started page is displayed.
3 On the Host page, enter the search criteria and click Search.
The host that meets the search criteria is displayed.
Search criteria can be the host name.
Set the host to the maintenance mode.
4 Click the name of the target host in the host list.
5 In the upper part of the page, click Operation and select Enter Maintenance.
A confirmation dialog box is displayed.
6 Select Migrate All VMs if VMs running on the host need to be migrated to other hosts.
7 Click OK.
An information dialog box is displayed.
8 Click OK.
The host is set to the maintenance mode.
To have the host exit the maintenance mode, click Operation in the upper part of the
page and select Exit Maintenance.

----End

3.8 Changing the Multipathing Type of a Host

Scenarios
On FusionCompute, change the multipathing type of a host that uses Huawei storage area
network (SAN) devices. The default multipathing type is Universal, which indicates that non-
Huawei storage devices are used.

Impact on the System

The host must be restarted to make the new setting take effect, but services on the host will be
temporarily interrupted during the restart.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 45

Prerequisites
Conditions
You have logged in to FusionCompute.

Procedure
Search for the target host.
1 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
2 In the navigation tree on the left, select the target site.
The Getting Started page is displayed.
3 On the Host page, enter the search criteria and click Search.
The host that meets the search criteria is displayed.
Search criteria can be the host name.
Migrate VMs on the host.
4 Click the name of the target host in the host list.
The Getting Started page is displayed.
5 On the VM page, migrate the VMs on the target host to other hosts.
For details, see VM Operation Management > Migrating a VM in the FusionCompute
V100R005C10 Virtual Machine Management Guide.
Change the multipathing type of the host.
6 In the navigation tree on the left, click the name of the target host.
7 Click Configure Storage Multipathing.
A dialog box is displayed, as shown in Figure 3-3.

Figure 3-3 Configure Storage Multipathing

8 Select the required multipathing type.

9 Click Restart Now.
An information dialog box is displayed.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 46

NOTE
This operation is not allowed when VMs exist on the host, try again after the VMs are migrated or
deleted.
10 Click OK.
The multipathing type of the host is changed.
----End

3.9 Setting Time Synchronization on a Host

Scenarios
On FusionCompute, configure an external clock source and the time synchronization
function for a host so that the host provides precise time by periodically synchronizing time
from the external clock source.

Impact on the System

This operation interrupts services on the host temporarily because the host service process
must be restarted after time synchronization is configured. Therefore, exercise caution when
performing this operation.

NOTE
If a host is set to the internal clock source, the configuration causes the restart of the service process on
the host. If more than 40 VMs run on the host, the service process restart will take a long time, resulting
in VMs that are running on the host to trigger the High Availability (HA) tasks. However, the VMs will
not be migrated to another host. After the service process restarts, the HA tasks will be automatically
canceled.

Prerequisites
Conditions
l You have logged in to FusionCompute.
l All the Network Time Protocol (NTP) servers use the same upper-layer clock source so
that the system times of the NTP servers are the same if multiple NTP servers are to be
deployed.
l The domain name server (DNS) is configured if time synchronization is configured using
a domain name. For details, see System Configuration > Configuring the DNS Server
in the FusionCompute V100R005C10 Configuration Management Guide.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 47

Search criteria can be the host name.

Set the time synchronization function on the host.
4 Click the name of the target host in the host list.
5 Click Configure Time Synchronization.
A dialog box is displayed, as shown in Figure 3-4.

Figure 3-4 Configure Time Synchronization

NOTICE
If multiple NTP servers are to be deployed, ensure that all the NTP servers use the same
upper-layer clock source so that the system times of the NTP servers are the same.

Configure the host time synchronization parameters, including NTP server and
Synchronization interval(s).
Note the configuration requirements for the following parameters:
– NTP server: specifies the IP address or domain name of an NTP server. Enter one
to three IP addresses of the NTP servers. If time synchronization is configured
using a domain name, the DNS server must be configured in advance.
– Synchronization interval(s): specifies the time synchronization duration.
7 Click OK.
A confirmation dialog box is displayed.
8 Click OK.
An information dialog box is displayed.
9 Click OK.
The time synchronization function is configured on the host.
----End

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 48

3.10 Performing Forcible Time Synchronization

Scenarios
On FusionCompute, perform forcible time synchronization to ensure that the system time on
the host is consistent with that on the Network Time Protocol (NTP) server. Perform this
operation when the NTP server time is changed or the system time on the host is inconsistent
with that on the NTP server.

Impact on the System

Forcible time synchronization restarts service processes. Therefore, you are advised to
perform this operation when the service traffic is low.

Prerequisites
Conditions
You have logged in to FusionCompute.
Data
You have obtained the name of the host on which the system time needs to be synchronized.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 49

Forcible time synchronization is performed on the host.

----End

3.11 Configuring Host BMC Parameters

Scenarios
On the FusionCompute, configure Baseboard Management Controller (BMC) IP address,
username, and password for a host so that the system can power on or power off it when
scheduling resources.

Prerequisites
Conditions
You have logged in to the FusionCompute.

Procedure
Step 1 On the FusionCompute, choose Computing Pool.

Step 2 In the navigation tree on the left, choose Site name > Cluster name > Host name.
The Getting Started page is displayed.

Step 3 Click BMC Settings.

A dialog box is displayed, as shown in Figure 3-5.

Figure 3-5 BMC Settings

Step 4 Enter the BMC parameters, including the IP address, username, and password for logging into
the BMC.

Step 5 Click OK.

A dialog box is displayed.

Step 6 Click OK.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 50

The BMC parameters are configured for the host.

----End

3.12 Sharing the GPU on a Host

Scenarios
On FusionCompute, host memory sharing is employed to share the Graphic Processing Unit
(GPU) of a host so that the GPU is accessible to multiple VMs.
GPU sharing is supported only by the FusionCloud Desktop Solution.

NOTE
The antivirus virtualization function and GPU sharing function cannot be enabled on a host at the same
time.

Impact on the System

The host must be restarted on the FusionCompute web client to make the sharing settings
take effect. During host restart, however, services on the host will be interrupted temporarily.

Prerequisites
Conditions
You have logged in to FusionCompute.

Procedure
Search for the target host.
1 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
2 In the navigation tree on the left, select the target site.
The Getting Started page is displayed.
3 On the Host page, enter the search criteria and click Search.
The host that meets the search criteria is displayed.
Search criteria can be Name.
Share the host GPU.
4 Click the name of the target host in the host list.
5 On the Getting Started page, click Configure GPU Sharing.
A dialog box is displayed.
6 Select Enable GPU sharing.
7 From the drop-down list, select the number of VMs that can access the shared host GPU.
– Number of GPU clients: specifies the number of the VMs that can access the shared
host GPU.
– Shared memory for data sending by GPU clients (MB): specifies the memory size
of the GPU client VM that sends data to the GPU service VM.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 51

– Shared memory for data receiving by GPU clients (MB): specifies the memory size
of the GPU service VM that sends data to the GPU client VM.
NOTE
Note the algorithm when configuring the number of the VMs: (Shared memory for data sending by
GPU clients + Shared memory for data receiving by GPU clients) x Number of GPU clients ≤
1552 MB.
8 Click OK.
An information dialog box is displayed.
9 Click OK.
The host GPU is shared.
The sharing settings can take effect only after the host is restarted on the
FusionCompute web client.
To modify the sharing settings, refer to 5 to 9. The modification takes effect upon restart.
Ensure that the number of VMs that can access the shared host GPU must be greater than
or equal to the number of available GPU clients.

----End

3.13 Configuring the Host GPU Mode

Scenarios
On FusionCompute, configure the general process unit (GPU) mode for a GPU on a host.
The modes include passthrough mode and virtualization mode.

A host GPU in passthrough mode can be attached to only one VM. A host GPU in
virtualization mode can be shared by multiple VMs if the GPU supports virtualization.

GPU modes can be configured only when FusionAccess is in use.

Prerequisites
Conditions
l You have logged in to FusionCompute.
l The host GPU has not been attached to any VM.
l The GPU model is Nvidia Quadro 2000, Nvidia Quadro 4000, Nvidia GRID K1, Nvidia
GRID K2, Nvidia Quadro K2000, or Nvidia Quadro K4000 if you determine to set the
GPU to passthrough mode.
l The GPU model is NVIDIA GRID K1 or NVIDIA GRID K2 if you determine to set the
GPU to virtualization mode.
l You have obtained the host GPU driver if you determine to set the GPU to virtualization
mode. To obtain the GPU driver, contact Technical Support.

Procedure
Search for the target host.
1 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 52

2 In the navigation tree on the left, choose the site containing the target host.
The Getting Started page is displayed.
3 On the Host page, enter the search criteria and click Search.
The hosts that meet the search criteria are displayed.
Search criteria can be the host name.
Install the host GPU driver.
4 Check the GPU mode to be set.
– If it is the passthrough mode, go to 11.
– If it is the virtualization mode, go to 5.
5 Obtain the IP address of the target host in the list.
6 Decompress the obtained GPU driver package NVIDIA-vgx-uvp-xxx-xxx.x86_64.rar.
NVIDIA-vgx-uvp-xxx-xxx.x86_64.rpm is obtained.
7 Use WinSCP to upload NVIDIA-vgx-uvp-xxx-xxx.x86_64.rpm to the /home directory
on the host.
8 Use PuTTY to log in to the host as user gandalf and run the su - root command to
switch to user root.
9 Run the cd /home command to switch to the /home directory.
10 Run the following command to install the host GPU driver:
rpm -ivh NVIDIA-vgx-uvp-xxx-xxx.rpm
Configure the GPU mode.
11 Click the name of the target host in the host list on FusionCompute, click the Device
tab, and click GPU.
12 Locate the row containing the target GPU, and click Modify in the Operation column.
A dialog box is displayed, as shown in Figure 3-6.

Figure 3-6 Modify GPU Mode

13 From the drop-down list, select a GPU mode.

– If the host GPU does not support virtualization, you can only select
Passthrough(1:1).
– If the host GPU supports virtualization (only NVIDIA GRID K1 and NVIDIA
GRID K2 support GPU virtualization), you can set any of the modes listed in Table
3-1 besides Passthrough(1:1).
– If the performance of a virtual GPU (vGPU) provided by a host GPU does not meet
service requirements, set the GPU mode to Passthrough(1:1).

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 53

Table 3-1 vGPU modes

Physical GPU vGPU vGPU Video
vGPU Mode
Model Resolution RAM

NVIDIA GRID K1 k100(1:8) 1920x1200 256 MB

k120q(1:8) 2560x1600 416 MB

k140q(1:4) 2560x1600 896 MB

NVIDIA GRID K2 k200(1:8) 1920x1200 256 MB

k220q(1:8) 2560x1600 416 MB

k240q(1:4) 2560x1600 896 MB

k260q(1:2) 2560x1600 1856 MB

14 Click OK.
An information dialog box is displayed.
15 Click OK.
The host GPU mode is set.
To change the GPU mode of other GPUs on the host, see 12 to 15.
----End

3.14 Enabling or Disabling LLDP for a Host

Scenarios
FusionCompute supports the host Link Layer Discovery Protocol (LLDP) service, which
enables a host to report the host name, port names, MAC address, and IP address to the
switch. This service is disabled by default. You can use the FusionCompute web client to
enable or disable the LLDP service for a host.

Prerequisites
Conditions
l You have obtained the target host name and the port name.
NOTE

The LLDP service cannot be enabled for a host that uses intelligent network interface cards
(iNICs).
l You have logged in to FusionCompute.

Procedure

1 On FusionCompute, click Computing Pool.

The Computing Pool page is displayed.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 54

2 In the navigation tree on the left, select the site, cluster, and host.
The Getting Started page is displayed.
3 Click Device.
4 Select PCI Device.
5 Determine whether to enable or disable LLDP for the host.
– To enable LLDP, go to 6.
– To disable LLDP, go to 9.
6 Locate the row that contains the target host port and click Enable LLDP.
A dialog box is displayed, asking you whether to enable the LLDP service.
7 Click OK.
A dialog box is displayed.
8 Click OK.
No further action is required.
9 Locate the row that contains the target host port and click Disable LLDP.
A dialog box is displayed, asking you whether to disable the LLDP service.
10 Click OK.
A dialog box is displayed.
11 Click OK.
No further action is required.

----End

3.15 Enabling the Host Antivirus Function

Scenarios
On FusionCompute, enable the antivirus function for a host to provide VMs on it with virus
scan, removal, and real-time monitoring services. For details about the antivirus function and
its deployment plans, see VM Antivirus Management in the FusionCompute V100R005C10
Virtual Machine Management Guide.

NOTE
The antivirus virtualization function and GPU sharing function cannot be enabled on a host at the same
time.

Impact on the System

The host must be restarted to make the settings take effect. During the restart process,
however, services on the host will be interrupted temporarily.

Prerequisites
Conditions
You have logged in to FusionCompute.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 55

Procedure
Step 1 On FusionCompute, click Computing Pool.

Step 2 In the navigation tree on the left, select the site, cluster, and host.
The Getting Started page is displayed.

Step 3 Click Antivirus.

A dialog box is displayed, as shown in Figure 3-7.

Figure 3-7 Antivirus

Step 4 Select on.

Step 5 Enter the number of the required secure user VMs on the host.
NOTE

Each host supports a maximum of 50 secure user VMs. However, the number of supported VMs on a
host varies depending on the antivirus software. If the number exceeds the upper limit of the software,
the antivirus functions become invalid.

Step 6 Click OK.

A dialog box is displayed.

Step 7 Click OK.

Step 8 Click Restart and click OK twice.

Restart the host for the configuration to take effect.

----End

3.16 Moving a Host

Scenarios
On the FusionCompute, move a host:
l From a site to a cluster
l From one cluster to another

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 56

l From a cluster to a site

Impact on the System

If a host is moved from a cluster to a site and the VMs that run on the host use shared storage
that is not used by other hosts in the cluster, you must bind the VMs back to the host after the
host is moved before you can start the VMs.

Prerequisites
Conditions
l You have logged in to the FusionCompute.
l You have stopped or migrated the VMs on the host.
l You have migrated or deleted the VMs using the local storage of the host.
l The CPU generation of the host to be moved to the IMC-enabled cluster is the same as or
later than the IMC mode configured for the cluster. Execute Disable Bit function, which
is also known as the No eXecute bit (NX) or eXecute Disable (ND) function, has been
enabled in the advanced CPU options for the BIOS of the host.

Procedure
Step 1 On the FusionCompute, choose Computing Pool.
The Computing Pool page is displayed.
Step 2 In the navigation tree on the left, choose Site name > Cluster name > Host name.
The Getting Started page is displayed.
Step 3 Right-click the host to be moved, and choose Move.
A dialog box is displayed, as shown in Figure 3-8.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 57

Figure 3-8 Move Host

Step 4 Select the destination to which the host is moved.

You may choose to move the host to another cluster or site.
Step 5 Click OK.
A dialog box is displayed.
Step 6 Click OK.
The host is moved from the cluster or site.

----End

3.17 Removing a Host

Scenarios
On the FusionCompute, remove a host from a cluster. After the host is removed, all
resources provided by the host will be deducted from the total resources provided by the
cluster.
You can remove a host in Remove or Forcibly Remove mode. In normal cases, Remove is
recommended for removing a host safely. If the host is faulty and cannot be removed in
Remove mode, you can remove it forcibly in Forcibly Remove mode.
If a host is associated with a shared storage resource that is associated with only this host, the
host cannot be removed.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 58

Impact on the System

l The host is disconnected from the system after being removed from the cluster.
l If the host is removed forcibly, it may fail to be added again. In this case, reinstall the
operating system (OS) for the host and then add it to the cluster.

Prerequisites
Conditions
l You have logged in to the FusionCompute.
l If the host uses FusionStorage, the host has been deleted from the FusionStorage
management system before being removed from the cluster. This is to prevent host
removal from adversely affecting the FusionStorage system. For details, see the
FusionStorage V100R003C30SPC200 Capacity Adjustment Guide.
l The following conditions are met for removing a host in Remove mode:
– The VMs on the host have been deleted or migrated to other hosts.
– The host upstream link connected to the DVS has been removed. For details, see
Upstream Link Group Management > Deleting an Upstream Link in the
FusionCompute V100R005C10 Network Management Guide.
– The host has been disassociated from the associated data store and disconnected
from all storage resources.
– Data stores whose type is Local on the host have been destroyed.
– Heartbeat Communication Between the Host and VRM Interrupted is not
generated for the target host.
l The following conditions are met for removing a host in Forcibly Remove mode:
– The host is in the fault or poweroff state.
– The VMs running on the host have been stopped forcibly.
– If the host is associated with a shared storage resource, the shard storage resource
has been associated with multiple hosts including the one to be removed.
Data
The following data is available:
l The name of the host to be removed
l The name of the cluster from which the host is to be removed

Procedure

1 On the FusionCompute, choose Computing Pool.

The Computing Pool page is displayed.
2 In the navigation tree on the left, choose Site name > Cluster name > Host name.
The Getting Started page is displayed.
3 Right-click the host to be removed and choose Remove or Forcibly Remove as
required.
A dialog box is displayed.
4 Click OK.
A dialog box is displayed.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 59

5 Click OK.
The host is removed from the cluster.
Follow-up Procedure
Power off the host immediately after removing it in Forcibly Remove mode. Otherwise,
some functions, such as storage device attaching, on other normal hosts in the same site may
fail.
----End

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 60

4 System Port Management

About This Chapter

4.1 Binding Network Ports

4.2 Adding a Storage Port
4.3 Adding Service Management Ports (Required Only for Virtualized SAN Storage)
4.4 Querying Routes of a System Port
4.5 Modifying Attributes of a System Port
4.6 Deleting a Storage Port
4.7 Deleting a Service Management Port

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 61

4.1 Binding Network Ports

Scenarios
On FusionCompute, bind network ports on a host to improve network reliability.

NOTE
If the host uses intelligent network interface cards (iNICs), bind the uplink network ports on the host together.
Otherwise, the broadcast suppression function of the port group may be adversely affected.

Prerequisites
Conditions
l You have logged in to FusionCompute.
l The host has been added to a cluster.

Procedure
Determine the method of binding network ports.
1 Determine the method of binding network ports.
– To bind network ports in batches, go to 10.
This method is recommended when multiple bound ports are to be added to hosts or
the hosts are large in number.
– To bind network ports one by one, go to 2.
This method is recommended when the hosts are small in number and each host
requires a few bound ports.
Bind network ports one by one.
2 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
3 In the navigation tree on the left, select the site, the cluster, and the host.
The Getting Started page is displayed.
4 Choose Configuration > System Port > Bind Network Port.
The Bind Network Port page is displayed, as shown in Figure 4-1.

Figure 4-1 Bind Network Port

5 In the Network Port list, select the physical network ports to be bound.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 62

6 In the middle of the page, set Name and Binding Mode for the network ports.

NOTICE
– In all load sharing modes, aggregation must be configured on the switch to which
network ports are connected, that is, the ports to be bound must be configured on the
same Eth-trunk port on the switch. Otherwise, network exception may occur.
– In the Link Aggregation Control Protocol (LACP) mode, create an Eth-trunk in
LACP mode on the switch to which network ports are connected, configure ports to
be bound on the same Eth-trunk, and enable the bridge protocol data unit (BPTU)
protocol packet forwarding function on the Eth-trunk. For example, if the switch is
Huawei S5300, run the following commands:
<S5352_01>sys
[S5352_01]interface Eth-Trunk x
[S5352_01-Eth-Trunkx]mode lacp-static
[S5352_01-Eth-Trunkx]bpdu enable
For details about how to configure port aggregation on a switch, see the switch user
guide.

The following binding modes are available for common network interface cards (NICs):
– Active-backup: applies to scenarios where two network ports are to be bound. This
mode provides high reliability. The bandwidth of the bound port in this mode equals
to that of a member port.
– Round-robin: applies to scenarios where two or more network ports are to be
bound. The bandwidth of the bound port in this mode is higher than that of a
member port, because the member ports share workloads in sequence.
This mode may result in data packet disorder because traffic is evenly sent to each
port. Therefore, MAC address based load balancing prevails over Polling in load
sharing modes.
– IP address and port-based load balancing: applies to scenarios where two or
more network ports are to be bound. The bandwidth of the bound port in this mode
is higher than that of a member port, because the member ports share workloads
based on the source-destination-port-based load sharing algorithm.
Source-destination-port-based load balancing algorithm: When the packets
contain IP addresses and ports, the member ports share loads based on the source
and destination IP addresses, ports, and MAC addresses. When the packets contain
IP addresses, the member ports share loads based on the IP addresses and MAC
addresses. When the packets contain only MAC addresses, the member ports share
loads based on the MAC addresses.
This mode is recommended when the virtual extensible LAN (VXLAN) function is
enabled. This mode allows network traffic to be evenly distributed based on the
source and destination port information in the packets.
– MAC address-based load balancing: applies to scenarios where two or more
network ports are to be bound. The bandwidth of the bound port in this mode is
higher than that of a member port, because the member ports share workloads based
on the MAC addresses of the source and destination ports.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 63

This mode is recommended when most network traffic is on the layer 2 network.
This mode allows network traffic to be evenly distributed based on the MAC
addresses.
– MAC address-based LACP: This mode is developed based on the MAC address
based load balancing mode. In MAC address-based LACP mode, the bound port
can automatically detect faults on the link layer and trigger a switchover if a link
fails using the LACP protocol.
– IP address-based LACP: applies to scenarios where two or more network ports are
to be bound. The bandwidth of the bound port in this mode is higher than that of a
member port, because the member ports share workloads based on the source-
destination-IP-address-based load sharing algorithm. When the packets contain IP
addresses, the member ports share loads based on the IP addresses and MAC
addresses. When the packets contain only MAC addresses, the member ports share
loads based on the MAC addresses. In this mode, the bound port can also
automatically detect faults on the link layer and trigger a switchover if a link fails
using the LACP protocol.
This mode is recommended when most network traffic goes across layer 2 and layer
3 networks.
The following binding modes are available for intelligent network interface cards
(iNICs):
– Active-backup: applies to scenarios where two network ports are to be bound. This
mode provides high reliability. The bandwidth of the bound port in this mode equals
to that of a member port.
– Source MAC address-based load balancing: applies to scenarios where two or
more network ports are to be bound. The bandwidth of the bound port in this mode
is higher than that of a member port, because the member ports share workloads
based on the MAC address of the source port.
– Destination MAC address-based load balancing: applies to scenarios where two
or more network ports are to be bound. The bandwidth of the bound port in this
mode is higher than that of a member port, because the member ports share
workloads based on the MAC address of the destination port.
This mode is recommended when most network traffic is on the layer 2 network.
This mode allows network traffic to be evenly distributed based on the MAC
addresses.
– Source IP address-based load balancing: applies to scenarios where two or more
network ports are to be bound. The bandwidth of the bound port in this mode is
higher than that of a member port, because the member ports share workloads based
on the IP address of the source port.
– Destination IP address-based load balancing: applies to scenarios where two or
more network ports are to be bound. The bandwidth of the bound port in this mode
is higher than that of a member port, because the member ports share workloads
based on the IP address of the destination port.
This mode is recommended when most network traffic is on the layer 3 network.
This mode allows network traffic to be evenly distributed based on the destination
IP addresses.
7 Click Bind.
An information dialog box is displayed.
8 Click OK.
The Bind Network Port page is displayed.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 64

To change the binding mode of a bound port, locate the row that contains the bound port,
click Operation, and click Modifying binding mode. You can change the binding mode
in the displayed dialog box.

NOTICE
– Switching between different load sharing modes or between different LACP modes
interrupts network communication of the bound network port for 2 or 3 seconds.
– If the binding mode is changed from the active/standby mode to load sharing mode,
port aggregation must be configured on the switch to which network ports are
connected. If the binding mode is changed from the load sharing mode to active/
standby mode, the aggregation configured on the switch must be canceled.
Otherwise, network exception may occur.
– If the binding mode is changed from the LACP mode to another mode, port
configuration must be changed on the switch to which network ports are connected. If
the binding mode is changed from another mode to the LACP mode, port aggregation
in LACP mode on the switch. Otherwise, network exception may occur.
Configuration operations on the switch may interrupt the network communication. After
the configurations are complete, the network communication is automatically restored. If
the network communication is not restored, perform either of the following methods to
troubleshoot the network:
– Ping the destination IP address from the switch to trigger a MAX table update.
– Select a member port in port aggregation, disable other ports on the switch, change
the binding mode, and enable those ports.

9 Click OK.
The network ports on the host are bound.
After this step is complete, no further action is required.
Bind network ports in batches.
10 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
11 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.
12 In the Operation list of the cluster, click Bind Network Ports in Batches.
The Bind Network Ports in Batches page is displayed, as shown in Figure 4-2.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 65

Figure 4-2 Bind Network Ports in Batches

13 Click Obtain Template for Binding Network Ports in Batches.

The page for downloading templates is displayed.
14 Save the template on the local PC.
The default file name of the template is BindNetworkPortYYYY-MM-DD
HHMMSSXXX.xls.
15 Open the template file and set the network port parameters on the Config sheet based on
the information provided on the Host_BindNetworkPort page.
The parameters include Host IP Address, Host ID, Network Port Name, Network
Port IDs, Bound Network Port Name, and Binding Mode.
16 Save and close the template file.
17 Click Browse on the right of Import template file on the Bind Network Ports in
Batches page.
A dialog box is displayed.
18 Select the configured template file and click Open.
19 Click OK.
An information dialog box is displayed.
20 Click OK.
You can choose System > Tasks and Logs > Task Center to view the task progress.
Add network ports in batches.
21 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
22 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.
23 In the Operation list of the cluster, click Bind Network Ports in Batches.
The Bind Network Ports in Batches page is displayed, as shown in Figure 4-3.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 66

Figure 4-3 Bind Network Ports in Batches

24 Click Obtain Template for Add Network Ports in Batches.

The page for downloading templates is displayed.
25 Save the template on the local PC.
The default file name of the template is AddPorttoBoundYYYY-MM-DD
HHMMSSXXX.xls.
26 Open the template file and set the network port parameters on the Config sheet based on
the information provided on the Host_AddNetworkPort page.
The parameters include Host IP Address, Host ID, Network Port Name, Network
Port ID, and IDs of Network Ports to Be Bound.
27 Save and close the template file.
28 Click Browse on the right of Import template file on the Bind Network Ports in
Batches page.
A dialog box is displayed.
29 Select the configured template file and click Open.
30 Click OK.
An information dialog box is displayed.
31 Click OK.
You can choose System > Tasks and Logs > Task Center to view the task progress.

----End

Additional Information
Related Tasks
Removing a Default Network Port from a Bound Port
The default network port is the first port added to the bound port. To remove the default port,
unbind the bound port and then bind the other non-default member ports. If the bound port is
used by any services, you must remove the services first.
1. Locate the row that contains the bound port, choose More > Unbind Port, and check
whether The network port aggregation is in use is displayed.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 67

– If yes, the bound port is in use. Go to 2.

– If no, go to 7.
2. Switch to the page for the host, choose Configuration > System Port, and check
whether the bound port is used as a storage port or a service management port or not.
– If yes, go to 3.
– If no, go to 4.
3. Migrate all the VMs from the host to other hosts or stop all the VMs on the host,
disassociate all the virtualized SAN storage from the host, and delete the storage port or
service management port from the bound port.
4. In Network Pool, select the uplink of the DVS and check whether the bound port is used
by the uplink.
– If yes, go to 5.
– If no, go to 6.
5. Delete the uplink. If the uplink is used by VMs, migrate the VMs to other hosts or delete
the VM NICs, and then delete the uplink.
6. Locate the row that contains the bound port and choose More > Unbind Port.
7. Bind the other member ports of the bound port other than the default port.
8. If you remove the services on the original bound port before you unbind the port, restore
the services on the newly bound port, for example, creating the system interface on the
bound port, associating with the uplink, and migrating the original VMs back to the host.

4.2 Adding a Storage Port

Scenarios
On FusionCompute, add a storage port to a host to enable the host to communicate with
storage devices. If you add multiple storage ports to the host, the multipathing function can be
implemented, that is, the host data can be transferred to storage devices through multiple
paths.

A host can have a maximum of four storage ports added.

The host does not need any storage port to connect to a local hard disk, Fibre Channel storage
area network (FC SAN) device, or a local RAM disk.

Prerequisites
Conditions
l You have logged in to FusionCompute.
l The host has been added to a cluster.
l Operations provided in Binding Network Ports have been performed on the host if the
host uses multiple network ports to connect to a storage device.

Procedure
Determine the method of adding storage ports to hosts.
1 Determine the method of adding storage ports to hosts.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 68

– To add storage ports to hosts in batches, go to 12.

This method is recommended when multiple storage ports are to be added a large
number of hosts or to hosts that use the multipathing function.
– To add storage ports to hosts one by one, go to 2.
This method is recommended when the hosts are small in number or do not use the
multipathing function.
Manually add storage ports to a host.
2 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
3 In the navigation tree on the left, select the site, the cluster, and the host.
The Getting Started page is displayed.
4 Choose Configuration > System Port > Add Storage Port.
The Add Storage Port page is displayed, as shown in Figure 4-4.

Figure 4-4 Add Storage Port

5 Select the network port to which the storage plane network interface card (NIC)
connects, and click Next.
PORTX indicates network port ethX on the host. To identify the ports on common
Huawei servers, see How to Identify Server Ports.
The Connection Settings page is displayed, as shown in Figure 4-5.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 69

Figure 4-5 Connection Settings

6 Set the following parameters based on the network plan:

– Name
– Description
– IP address: Enter the storage port IP address. It must be an idle IP address in the
same network segment as the storage device IP address.
NOTE
Except the storage port IP address, this IP address and other IP addresses of the host must be
on different network segments.
n If the storage plane uses a layer 2 network, set it to an idle IP address that
communicates with the storage plane. For example, if the storage IP address of

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 70

the storage device is 172.20.100.100 and subnet mask is 255.255.0.0, set the IP
address to 172.20.XXX.XXX.
n If the storage plane uses a layer 3 network, set it to an IP address that
communicates with the storage IP address of the storage device.
– Subnet mask: Enter the subnet mask of the storage plane.
– VLAN ID: Enter the VLAN ID of the storage plane.
– Switching mode: specifies the data exchange mode of the storage plane. The value
can be Linux subinterface or OVS forwarding.
NOTE

n If FusionStorage is used, use the Linux subinterface switching mode.

n You can configure the switching mode only for common NICs.
n The Linux subinterface switching mode minimizes the mutual influence between the
storage plane and management and service planes. The VLAN of the network port
associated with the storage port is exclusively used by the storage port, and the VLAN
ID is 0. The management plane, service plane, and other storage ports on this network
port cannot use the VLAN.
n The OVS forwarding mode allows the storage port, management plane, service plane,
and VXLAN Tunnel End Point (VTEP) to share a VLAN.
– Router info: specifies the route information for layer 3 network communication.
NOTE
If the host storage plane uses intelligent NICs (iNICs), the layer 3 routed network cannot be
configured.
n Gateway: specifies the gateway of the storage ports. The gateway address and
the storage IP address must belong to the same network segment.
n Network destination: specifies the first IP address of the network segment
that the target storage device uses, for example, 192.168.0.0.
n NetMask: specifies the subnet mask of the subnet network segment.
– Outbound Traffic Shaping
n Average send bandwidth (Mbit/s): specifies the average number of bits per
second to allow across a port during a certain period of time. If a common NIC
is used, the port traffic remains close to the configured average bandwidth
when no burst of traffic occurs. If an iNIC is used, the average bandwidth is
equal to the minimum bandwidth when no congestion occurs on the network.
If the burst send size is set to a too small value, the network bandwidth
decreases.
n Peak send bandwidth (Mbit/s): specifies the maximum number of bits per
second to allow across a port when it is sending a burst of traffic. The peak
send bandwidth must be greater than or equal to the average send bandwidth.
A proper peak send bandwidth set for a service prevents network congestion
on other VM networks when the traffic of this service is too large. When an
iNIC is used, the peak send bandwidth is equal to the maximum bandwidth
after the burst of traffic disappears, and in the idle period, the bandwidth
remains around the peak send bandwidth.
n Burst send size (Mbits): specifies the maximum number of bytes to allow in a
burst.

If SAN devices are used and multipathing is required, configure multiple storage
interfaces for a single storage network port.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 71

For example, if a storage device has four storage paths, using VLAN4, VLAN5,
VLAN6, and VLAN7, respectively, the network port eth2 on the host is to be configured
to intercommunicate with VLAN4 and VLAN5, and eth3 to intercommunicate with
VLAN6 and VLAN7, configure storage interfaces VLAN4 and VLAN5 on eth2, and
VLAN6 and VLAN7 on eth3.
7 Click Next.
The Confirm page is displayed.
8 Ensure the all the information is correct and click Add.
A dialog box is displayed.
9 Check whether all the planned storage ports have been added.
– If yes, go to 11.
– If no, go to 10.
10 Click Continue, and perform 5 to 8 for each storage port to be added.
11 Click OK.
The storage ports are added to the host.
After this step is complete, no further action is required.
Add storage ports to hosts in batches.
12 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
13 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.
14 Choose Operation > Add Storage Ports in Batches on top of the page.
The Add Storage Ports in Batches page is displayed, as shown in Figure 4-6.

Figure 4-6 Add Storage Ports in Batches

15 Click Obtain the Add Storage Ports in Batches template file.

The page for downloading templates is displayed.
16 Download the template to the PC.
The default file name of the template is StorageNetworkYYYY-MM-DD
HHMMSSXXX.xls.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 72

17 Open the template, click the Host Storage Port sheet, locate the row that contains
information about storage ports to be added, and copy the information in the row to the
Config sheet.
PORTX indicates network port ethX on the host. To identify the ports on common
Huawei servers, see How to Identify Server Ports.
Storage port information includes Host IP Address, Host ID, Network Port Name, and
Network Port ID.
18 On the Config sheet, set the following parameters for the storage ports:
– Storage Port Name
– Storage Port Description
– Storage IP Address: Enter the storage port IP address. It must be an idle IP address
in the same network segment as the storage device IP address.
n If the storage plane uses a layer 2 network, set it to an idle IP address that
communicates with the storage plane. For example, if the storage IP address of
the storage device is 172.20.100.100 and subnet mask is 255.255.0.0, set the IP
address to 172.20.XXX.XXX.
n If the storage plane uses a layer 3 network, set it to an IP address that
communicates with the storage IP address of the storage device.
– Subnet Mask: Enter the subnet mask of the storage plane.
– VLAN ID: Enter the VLAN ID of the storage plane.
– Switching mode: specifies the data exchange mode of the storage plane. The value
can be Linux subinterface or OVS forwarding.
NOTE

n If FusionStorage is used, use the Linux subinterface switching mode.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 73

For details about the parameters, see the help sheet in the template.
19 Save and close the template.
20 Click Browse to the right of Import template file on the Add Storage Ports in Batches
page.
A dialog box is displayed.
21 Select the template and click Open.
22 Click OK.
An information dialog box is displayed.
23 Click OK.
You can choose System > Tasks and Logs > Task Center to view the task progress.
Follow-up Procedure
After the storage ports are added to the hosts, If the hosts use shared storage resources, add
and associate the storage resources to hosts, and then scan storage devices. For details, see
Adding Storage Resources to a Site, Associating Storage Resources with a Host and
Scanning Storage Devices in the FusionCompute V100R005C10 Storage Management
Guide.

----End

4.3 Adding Service Management Ports (Required Only for

Virtualized SAN Storage)
Scenarios
On FusionCompute, add service managements port to a host. A service management port is
used to carry traffic for a special service, such as VM live migration, host-based replication
disaster recovery (DR), or Virtual Image Management System (VIMS) heartbeat detection.

If a host is added to a cluster and does not have any service management port added, the
traffic for the special services and management services is carried by the management port by
default. After service management ports are added to a host and are enabled, the service
management ports carry and manage the traffic for the special services.

To enable the live migration or host-based replication function for a cluster, you are advised to
enable or disable the corresponding function for service management ports on all hosts in the
cluster.

A host supports a maximum of four service management ports. A host that uses intelligent
network interface cards (iNICs) does not support service management ports.

Prerequisites
Conditions
l You have logged in to FusionCompute.
l Hosts have been added to a cluster.
l You have enabled an independent network plane that carries the virtualized SAN storage
traffic if you want to add a service management port that manages the virtualized SAN
storage heartbeat traffic. For details, see Changing the Network Plane that Carries

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 74

Virtualized SAN Storage Traffic in the FusionCompute V100R005C10 Storage

Management Guide.

Procedure
Determine the method of adding service management ports to hosts.
1 Determine the method of adding service management ports to hosts.
– To add service management ports to hosts in batches, go to 10.
This method is recommended when the system has a large number of hosts.
– To add service management ports to hosts one by one, go to 2.
This method is recommended when the system has a small number of hosts.
Manually add a service management port to a host.

2 On FusionCompute, click Computing Pool.

3 In the navigation tree on the left, select the site, the cluster, and the host.
The Getting Started page is displayed.
4 Choose Configuration > System Port > Add Service Management Port.
The Add Service Management Port page is displayed, as shown in Figure 4-7.

Figure 4-7 Add Service Management Port

5 Select the host network port that connects to the service plane, and click Next.
PORTX indicates network port ethX on the host. For details about common network
ports on Huawei servers, see How to Identify Server Ports.
The Connection Settings page is displayed, as shown in Figure 4-8.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 75

Figure 4-8 Connection Settings

6 Set the following parameters based on the network plan:

– Name
– Description
– IP address: specifies the IP address of the service management port.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 76

NOTE
This IP address and IP addresses of other ports of the host must be on different network
segments.
– Subnet mask: specifies the subnet mask of the planned service management plane.
– VLAN ID: specifies the VLAN ID of the planned service management plane.
– Routing info: specifies the information about the route to the peer host. This
parameter is required when the service management port on the local host and the
service management ports on other hosts in the cluster do not belong to the same
network segment and the service management port on the local host is enabled.
n Gateway: specifies the gateway on the network segment to which the service
management port on the destination host belongs.
n Network destination: specifies the start IP address of the network segment to
which the service management port on the destination host belongs, for
example, 192.168.0.0.
n Netmask: specifies the subnet mask of the network segment to which the
service management port on the destination host belongs.
– Available Services
n Use this port for VM live migration: makes the service management port
carry the VM live migration traffic.
n Use this port for VM DR: makes the service management port carry the host-
based replication DR traffic.
n Use this port for virtualized SAN storage traffic: makes the service port
carry the VIMS heartbeat traffic.
– Outbound Traffic Shaping
n Average send bandwidth (Mbit/s): specifies the average number of bits per
second to allow across a port during a certain period of time.
If a common NIC is used, the port traffic remains close to the configured
average bandwidth when no burst of traffic occurs. If an iNIC is used, the
average bandwidth is equal to the minimum bandwidth when no congestion
occurs on the network. If the burst send size is set to a too small value, the
network bandwidth decreases.
n Peak send bandwidth (Mbit/s): specifies the maximum number of bits per
second to allow across a port when it is sending a burst of traffic.
The peak receive bandwidth must be greater than or equal to the average
receive bandwidth. A proper peak send bandwidth set for a service prevents
network congestion on other VM networks when the traffic of this service is
too large. When an iNIC is used, the peak send bandwidth is equal to the
maximum bandwidth after the burst of traffic disappears, and in the idle
period, the bandwidth remains around the peak receive bandwidth.
n Burst send size (Mbits): specifies the maximum number of bytes to allow in a
burst.
7 Click Next.
The Confirm page is displayed.
8 Confirm the information and click Add.
The Finish page is displayed.
9 Click Close.
You can view the task progress on the Task Tracing page.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 77

The service management port is added to the host.

Add service management ports to hosts in batches.
10 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
11 In the navigation tree on the left, select the site and cluster.
The Getting Started page is displayed.
12 In the Operation list in the upper corner, click Add Service Management Port in
Batches.
The Add Service Management Port in Batches page is displayed, as shown in Figure
4-9.

Figure 4-9 Add Service Management Port in Batches

13 Click Download Template.

The page for downloading templates is displayed.
14 Save the template on the local PC.
The default file name of the template is StorageNetworkYYYY-MM-DD
HHMMSSXXX.xls.
15 Open the template file, click the Host_ServiceManagement Port sheet, locate the row
that contains the information about service management ports to be added, and copy the
information in the row to the Config sheet.
PORTX indicates network port ethX on the host. For details about common network
ports on Huawei servers, see How to Identify Server Ports.
The information in the row includes Host IP Address, Host ID, Network Port Name,
and Network Port ID.
16 On the Config sheet, set the following parameters for the service management ports:
– Service Management Port Name
– Service Management Port Description
– IP Address: specifies the IP address of the service management port.
– Subnet Mask: specifies the subnet mask of the planned service management plane.
– VLAN ID: specifies the VLAN ID of the planned service management plane.
– Use this port for VM live migration: makes the service management port carry the
VM live migration traffic.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 78

– Use this port for VM DR: makes the service management port carry the host-based
replication DR traffic.
– Use this port for virtualized SAN storage traffic: makes the service management
port carry the VIMS heartbeat traffic.
– Outbound Traffic Shaping: specifies whether to limit the bandwidth on the port.
Select Enable or Disable from the drop-down menu.
– Average Send Bandwidth (Mbit/s)
– Peak Send Bandwidth (Mbit/s)
– Burst Send Size (Mbit)
For details about the parameters, see the help sheet in the template.
17 After information about all service management ports is configured, save and close the
template file.
18 Click Browse on the right of Import template on the Add Service Management Port
in Batches page.
A dialog box is displayed.
19 Select the configured template file and click Open.
20 Click OK.
An information dialog box is displayed.
21 Click OK.
You can choose System > Tasks and Logs > Task Center to view the task progress.
----End

4.4 Querying Routes of a System Port

Scenarios
On FusionCompute, query the information about routes of a system port, including the
service management port route, management port route, and storage port route.
For operations about how to modify the service management port route and storage port route,
see Modifying Attributes of a system port.

Prerequisites
Conditions
You have logged in to FusionCompute.

Procedure
Step 1 On FusionCompute, click Computing Pool.
The Computing Pool page is displayed.
Step 2 In the navigation tree on the left, select the site, cluster, and host.
The Getting Started page is displayed.
Step 3 Choose Configuration > System Port.
All the system ports provided by the host are displayed.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 79

Step 4 Locate the row that contains the system port and click .
All routes of the system port are displayed in the list, as shown in the following figure.

Figure 4-10 Routing Table

Step 5 Close the window after you view the information.

----End

4.5 Modifying Attributes of a System Port

Scenarios
On FusionCompute, modify the attributes of a system port on a host. The attributes that can
be modified include port name, description, IP address, storage port and service management
port routes (excluding the management port route), live migration, and outbound traffic
shaping.
Route information must be configured of the storage plane communications over the layer 3
network. Before disabling the storage port route, ensure that the layer 2 network for the
storage plane is available.
If a host uses a virtualized storage area network (SAN) device, the management port IP
address cannot be changed on the FusionCompute portal. You can change is manually. For
details, see Network Adjustment > Changing VRM and Host Management IP Addresses
in Virtualized SAN Storage Devices Scenario in the FusionCompute V100R005C10
Configuration Management Guide.

Impact on the System

Network configuration may change if the attributes of a system port are changed. Exercise
caution when modifying the attributes of a system port.

Prerequisites
Conditions

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 80

You have logged in to FusionCompute.

Step 4 Click for the system port.

A dialog box is displayed.
Step 5 Modify the attributes of the system port.
l The attributes of a management port include Name, Description, IP address, Subnet
mask, Gateway, Outbound Traffic Shaping, Average send bandwidth (Mbit/s), Peak
send bandwidth (Mbit/s), and Burst send size (Mbits).
l The attributes of a storage port include Name, Description, IP address, Subnet mask,
VLAM ID, Switching mode, Routing info, Outbound Traffic Shaping, Average send
bandwidth (Mbit/s), Peak send bandwidth (Mbit/s), and Burst send size (Mbits).
l The attributes of a service management port include Name, Description, IP address,
Subnet mask, VLAM ID, live migration, VM DR, Virtualized SAN storage traffic,
Outbound Traffic Shaping, Average send bandwidth (Mbit/s), Peak send bandwidth
(Mbit/s), and Burst send size (Mbits).
Route information must be configured of the storage plane communications over the layer 3
network. Before disabling the storage port route, ensure that the layer 2 network for the
storage plane is available.
If a host uses a virtualized storage area network (SAN) device, the management port IP
address cannot be changed on the FusionCompute portal. You can change is manually. For
details, see Network Adjustment > Changing VRM and Host Management IP Addresses
in Virtualized SAN Storage Devices Scenario in the FusionCompute V100R005C10
Configuration Management Guide.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 81

Figure 4-11 Modify management port

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 82

Figure 4-12 Modify storage port

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 83

Figure 4-13 Modify service management port

Step 6 Click OK after modification.

An information dialog box is displayed.
Step 7 Click OK.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 84

The port group attribute modification task is complete.

----End

4.6 Deleting a Storage Port

Scenarios
On the FusionCompute, delete a storage port on the host.

NOTE
The management port cannot be deleted.

Impact on the System

If a storage port is deleted, the VM is disconnected from the storage device.

Prerequisites
Conditions
You have logged in to the FusionCompute.

Procedure
Step 1 On the FusionCompute, choose Computing Pool.
The Computing Pool page is displayed.
Step 2 In the navigation tree on the left, choose Site name > Cluster name > Host name.
The Getting Started page is displayed.
Step 3 Choose Configuration > System Port.
All the system ports provided by the host are displayed.

Step 4 Click for the storage port to be deleted.

An information dialog box is displayed, asking you whether to delete the storage port.
Step 5 Click OK.
An information dialog box is displayed, indicating that the storage port is deleted.
Step 6 Click OK.
The storage port is deleted.

----End

4.7 Deleting a Service Management Port

Scenarios
On FusionCompute, delete a service management port from a host.

NOTE
The management port cannot be deleted.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 85

Impact on the System

Deleting a service management causes failures in in-progress migration tasks.

Prerequisites
Conditions
You have logged in to FusionCompute.

Step 4 Click for the service management port to be deleted.

An information dialog box is displayed, asking you whether to delete the port.
Step 5 Click OK.
An information dialog box is displayed, indicating that the port is deleted.
Step 6 Click OK.
The service management port is deleted.

----End

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 86

A Appendix

A.1 FAQ
A.2 Parameter Reference
A.3 Parameters in Advanced Settings for Computing Resource Scheduling

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 87

A.1 FAQ

A.1.1 How to Configure Static Routes for Storage Ports on a Host

Scenarios
Configure static routes for storage ports on a host if the storage plane requires layer 3
communication, so that the storage IP addresses of storage devices and host storage ports that
do not belong to the same network segment can communicate with each other.

Impact on the System

If an existing route is deleted from a storage device, the storage link will be disconnected and
VMs created using the storage device will not provide services.

Prerequisites
Conditions
l You have installed the host operating system (OS).
l You have obtained the password for user root of the host.
Data
Data preparation is not required for this operation.

Procedure
Log in to the host.
1 Use PuTTY to log in to the host.
Ensure that the management IP address and username gandalf are used to establish the
connection.

The default password for user gandalf to log in to the VRM VM is Huawei@CLOUD8.
2 Run the following command and enter the password of user root to switch to user root:

su - root
3 Run the following command to disable logout on timeout:
TMOUT=0
Configure static routes for storage ports on the host.
4 Run the following command to switch to the /opt/galax/eucalyptus/ecs_scripts
directory:
cd /opt/galax/eucalyptus/ecs_scripts/
5 Run the following command to open the route parameter file using the visual interface
(vi) editor:
vi routes_param
6 Press i to enter editing mode.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 88

7 Enter route parameters in the following format:

Storage plane network segment Subnet mask Gateway address
– Storage plane network segment: Enter the network segment to which the storage
device IP address belongs.
– Subnet mask: Enter the bits of the subnet mask of the storage device.
– Gateway address: Enter the gateway address of the host storage port.
If the storage device is set to multipathing mode, add multiple routes.
For example, if the storage device IP addresses are 172.20.10.10 and 172.30.10.10, the
subnet masks are both 16-bit 255.255.0.0, and the host storage ports are 172.25.1.123
and 172.35.1.123, run the following commands to configure the routes:
172.20.0.0 16 172.25.1.1
172.30.0.0 16 172.35.1.1
8 Press Esc, enter :wq, and press Enter to save the configuration and exit from the vi
editor.
9 Run the following command for the routes to take effect:
sh add_static_routes.sh
10 Run the following command to create a scheduled task for the routes:
echo "* * * * * root sh /opt/galax/eucalyptus/ecs_scripts/add_static_routes.sh
>> /dev/null 2>&1" >> /etc/crontab
11 Run the following command to query the host OS installation location:
cat /etc/uvp_version | grep sys_type
12 Determine the subsequent operation based on the command output.
– If the command output contains the the following information, the OS is installed
on a local disk on the host. No further action is required.
sys_type=hdisk

– If the command output contains the the following information, the OS is installed
on a USB flash drive on the host. Go to 13.
sys_type=udisk

13 Run the following command to back up the routes:

uvp_configer_bakup -u GE -m /opt/galax/eucalyptus/ecs_scripts/routes_param -
m /etc/crontab
----End

Verification
Run the route command to query all routes of the host:

Additional Information
Related Tasks
Delete routes from host storage ports.
Perform the following operations to delete routes from host storage ports:
1. Log in to the host and run the following command to delete the scheduled task for routes:
sed -i '/add_static_routes.sh/d' /etc/crontab
2. In the /opt/galax/eucalyptus/ecs_scripts directory, open the routes_param file using
the vi editor and delete unwanted routes.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 89

3. Run the following command to delete information about the routes that have taken
effect:
route del -net Storage plane network segment/Subnet mask gw Gateway address
For example, if the route information to be deleted is 172.20.0.0 16 172.25.1.1, run the
following command:
route del -net 172.20.0.0/16 gw 172.25.1.1
4. Query the host OS installation location. If the OS is installed on a USB flash drive, route
information backup is required. For details about how to back up the information, see 11
to 13.

A.1.2 How to Identify Server Ports

This section helps you identify the ports on typical Huawei servers.

Ports on the RH2285 Server

Figure A-1 shows the ports on the RH2285.

Figure A-1 Ports on the RH2285

Ports on the RH2288 V2 Server

Figure A-2 shows the ports on the RH2288 V2.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 90

Figure A-2 Ports on the RH2288 V2

eth4 eth2

eth5 eth3 eth0 eth1 BMC

PLR02C7001

Ports on the RH2488 V2 Server

Figure A-3 shows the ports on the RH2488 V2.

Figure A-3 Ports on the RH2488 V2

BMC eth1 eth2 eth4

PLR03C7001

eth0 eth3 eth5

Ports on the X6000 Server

Figure A-4 shows the ports on the X6000.
l Port 1-0 is port eth0 on the upper blade.
l Port 2-0 is port eth0 on the lower blade.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 91

Figure A-4 Ports on the X6000

Ports on the E6000 Server

Figure A-5 shows the position of the MM610 management module and the ports of the
management module.

NOTE

By default, MM1, located in the left of the subrack, works as the active management module of the
E6000, and MM2, in the right of the shelf, works as the standby management module. You can
determine the active and standby management modules based on the indicators on the front panel of the
management modules.
l After the active management module is powered on, the ACT indicator is steady green.
l After the standby management module is powered on, the ACT indicator blinks green at 0.5 Hz.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 92

Figure A-5 Management ports on the E6000

Figure A-6 shows the position of the DM management module.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 93

Figure A-6 DM module in the E6000 (back panel)

SLOT C1

SLOT C2
SLOT B1

SLOT A1

SLOT A2

SLOT B2
PWR4 PWR5 PWR6

DM PWR1 PWR2 PWR3

PLE01003

Figure A-7 shows the ports of the DM module.

Figure A-7 DM ports on the E6000

Figure A-8 shows port 23 of switch module A1 in the E6000. The NX112 switch module is
used as an example.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 94

Figure A-8 Ports on the E6000

If NX113 switch module is used in the E6000, connect the local computer to the port on the
switch. Ensure that the connected port on the switch and the host management plane belong to
the same VLAN.

A.1.3 Failed to Execute the Policy for Scheduling Computing

Resources in a Cluster
Symptom
After the policy for scheduling computing resources in a cluster is enabled, the cluster is in
the Scheduling state all the time and cannot return to the Balance state. In addition, no
historical scheduling records or migration suggestions are provided.

Possible Cause
l The system does not automatically migrate VMs that do not meet migration
requirements, for example:
– The Keep VMs together or Mutually exclusive VMs rule is configured for VMs.
– The dynamic resource scheduling (DRS) function is disabled for VMs.
– A device is mounted to one or more VMs so that the dynamic resource scheduler
(DRS) cannot migrate these VMs. Therefore, the system cannot implement load
balancing.
– The VMs are incompatible with the destination host to which the DRS migrates the
VMs.
– The VM is bound to its host.
– The memory size configured for the VM is greater than 4 GB.
l The system does not migrate VMs in the event that balanced CPU and memory usage
cannot be realized after the migration. For example:
– When CPU and memory is selected, some hosts have high CPU capacity usage,
and the others have high memory usage.
– When CPU or Memory is selected, migrating a VM cannot solve the problem of
unbalanced CPU and memory usage.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 95

A.1.4 Background Information

Storage Resources
The FusionCompute can use storage resources provided by dedicated storage devices or local
disks on hosts. Dedicated storage devices are connected to hosts through network cables or
fiber cables.

Data Store
A data store is a storage unit that is converted from a storage resource by FusionCompute.
After a data store is associated with a host, the data store can be used to create virtual disks
for VMs.
Raw device mapping (RDM) allows logical unit numbers (LUNs) on SAN devices can serve
as data stores without creating virtual disks. This technology applies to large disk capacity
scenarios, for example, the database server construction. RDM can be used only for VMs that
run certain operating systems (OSs). For details about the supported OS list, see
Compatibility. If RDM storage is used to deploy application cluster services, such as Oracle
RAC, it is recommended that you not use the VM snapshot creation function and not restore a
VM using a snapshot. If you use a snapshot to restore a VM, the application cluster service
may become faulty.
After storage resources are converted to data stores, the difference between virtual disks
created using different resources are hidden from VM OSs.
Storage resources that can be converted to data stores are:
l LUNs on SAN devices, including Internet Small Computer Systems Interface (iSCSI)
storage devices and fiber channel (FC) storage devices.
l File systems on NAS devices.
l Storage pools on FusionStorage
l Local hard disks on hosts
l Local RAM disk on hosts
Table A-1 shows the relationship between data resources, storage devices and data stores.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 96

Table A-1 The storage devices, storage resources, and data stores support for the
FusionCompute
Storage Storage Data Store Storage Space
Device Resource Required By Data
Store

Local N/A Local (Non- Local: ≥2 GB

virtualization) Local VIMS: [2 GB, 16
Local VIMS TB)
(Virtualization) NOTE
l The remaining space on
the local disk where the
host OS is installed can
be added as a data store
of the local storage
type. If the space of the
disk is greater than 2
TB, the system
identifies the disk only
as a 2 TB disk during
the host OS installation
process. Therefore, the
remaining space on this
disk after the OS
installation is less than
2 TB. Other local disks
on the host are not
affected by the OS
installation. Therefore,
they can provide all of
their space.
l If local hard disks are
to be added as data
stores, it is
recommended that you
configure the disks in
RAID 1. If customers
have other requirements
on RAID levels,
configure the disks
based on customer
requirements.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 97

Storage Storage Data Store Storage Space

Device Resource Required By Data
Store

SAN device (IP SAN SAN (Non-virtualization) SAN: ≥ 2 GB

SAN or FC VIMS (Virtualization) VIMS: [5 GB, 64 TB]
SAN)
RDM (Raw device NOTE
mapping) After a logical unit number
(LUN) on virtualized SAN
storage (VIMS) is added as
a data store, the Virtual
Image Management System
(VIMS) takes up some
storage space. Therefore,
the actual available storage
space of the data store is
less than that of the LUN.
You can log in to the
FusionCompute web client
and choose Storage Pool >
Data Store > Summary >
Resource Statistics to view
detailed storage space used
by VIMS.
RDM: ≥ 2 GB

Advanced SAN Advanced Advanced SAN device N/A

device SAN device

NAS NAS NAS

FusionStorage FusionStorage FusionStorage

Local RAM N/A Local RAM disk (non- Local RAM disk: [16 GB,
disk virtualization) 512 GB]

Figure A-9 Data store association model

Data store association
RDM

VMs
File systems
Virtual disks
NAS devices

Hosts LUN
Data stores LUN

Local disks on hosts

SAN devices

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 98

Storage Port
A storage port on a host connects the host to a storage device. One physical NIC or a group of
physical NICs that are bound together on the host can be set as a storage port.
If iSCSI storage devices are used, two physical NICs on a host can be connected to multiple
storage NICs on the storage devices, working in multipathing mode. Binding of physical
NICs is not required in this mode.
If NAS devices are to be used, you are advised to bind the storage plane NICs in active/
standby mode and set the storage port to connect to the NAS devices to enhance reliability.

iSCSI Storage
An iSCSI storage device is connected to a host through network cables. The host accesses the
storage device using the TCP/IP protocol.
To ensure efficient access to an iSCSI storage device, configure an iSCSI initiator using the
world wide name (WWN) generated after the storage device is associated to the host.
Typical iSCSI storage devices include IP SAN devices and OceanStor 18000 series storage
devices.

FC Storage
An FC storage device is connected to the FC host bus adapter (HBA) on a host through
optical cables, which provide high data transmission rate.
To ensure efficient access to an FC storage device, configure an FC initiator using the WWN
generated after the storage device is connected to the host FC HBA.
Typical FC storage devices include FC SAN devices and OceanStor 18000 series storage
devices.

Multipathing
Multipathing is a storage access mechanism that provides more than one physical path to
connect a network storage device to one or more host network interface cards (NICs),
enabling load sharing for data flows and thereby enhancing reliability for storage access.
Usually, multipathing is supported by storage devices using iSCSI and Fibre Channel (FC),
such as IP SAN devices, FC SAN devices, and OceanStor 18000 series storage devices.
Multipathing supports the Huawei and universal multipathing modes. If the universal
multipathing mode is used and the VM uses raw device mapped disks, the MSCS cluster
cannot be deployed.for Windows Server OS. However, you can deploy the iSCSI network on
the VM to deploy the MSCS cluster.

Sample Multipathing Networking

This section uses Huawei OceanStor S5500T in Figure A-10 as an example to describe the
multipathing networking. The storage device in Figure A-10 has two controllers, A and B.
Each controller controls four storage NICs. The four NICs are assigned four VLANs, which
constitute the storage plane.
On a host, two NICs connect to the storage plane. Each host NIC communicates with two
storage NICs. In this case, two storage ports are required to be configured for each host NIC.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 99

Figure A-10 Sample multipathing networking

IP SAN

VLAN 4 Controller A
172.20.10.10
Host
VLAN 5
172.30.10.10
VLAN 6
Eth2 172.40.10.10
172.20.100.100
172.30.100.100 VLAN 7
172.50.10.10

Eth3 VLAN 4 Controller B

172.20.10.11
172.40.100.100
172.50.100.100 VLAN 5
172.30.10.11
VLAN 6
172.40.10.11
VLAN 7
172.50.10.11

The IP addresses shown in the figure are only used as examples.

NAS Storage
NAS storage devices use the network file sharing (NFS) protocol to provide shared folders
over a network.
A NAS storage device is connected to a host through network cables. A host accesses the
storage device using TCP/IP.

Local Storage
Disks on hosts provide local storage resources.
The FusionCompute can identify the following local storage resources:
l Free space on the disk on which the host operating system is installed
NOTE

The remaining space on the local disk where the host OS is installed can be added as a data store
of the local storage type. If the space of the disk is greater than 2 TB, the system identifies the disk
only as a 2 TB disk during the host OS installation process. Therefore, the remaining space on this
disk after the OS installation is less than 2 TB. Other local disks on the host are not affected by the
OS installation. Therefore, they can provide all of their space.
l Bare disks or unpartitioned redundant array of independent disks (RAID) on the host
Local storage can be provided only to the host housing the disks.

Local RAM Disk

A local random access memory (RAM) disk is generated from memory resources on a host
and provides storage resources for VMs only on the host.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 100

For details about adding local RAM disks on a host, see Creating Local RAM Disks on a
Host.

A.1.5 Compatibility
For details about the compatibility for servers, I/O devices, storage devices. and operating
systems (OSs), log in to compatibility check assistant.

Table A-2 lists OSs that support the PVSCSI.

Table A-2 OSs that support PVSCSI

Type OS

Windows Windows 2012 DataCenter 64-bit

Windows 2012 Standard 64-bit

Windows 2012 Standard R2 64-bit

Windows 2012 Essentials R2 64-bit

Windows 2012 DataCenter R2 64-bit

Windows Server 2008 Enterprise R2 64-bit

Windows Server 2008 Enterprise R2 SP1 64-bit

Windows Server 2008 Standard R2 64-bit

Windows Server 2008 Standard R2 SP1 64-bit

Windows Server 2008 WEB R2 64-bit

Windows Server 2008 DataCenter R2 64-bit

Windows Server 2008 DataCenter R2 SP1 64-bit

Windows Server 2008 DataCenter SP2 32-bit

Windows Server 2008 Enterprise SP1 32/64-bit

Windows Server 2008 Enterprise SP2 32/64-bit

Windows Server 2008 Standard SP1 32/64-bit

Windows Server 2008 Standard SP2 32/64-bit

Windows Server 2003 Standard R2 SP2 64-bit

Windows Server 2003 Enterprise R2 SP2 64-bit

Windows Server 2003 DataCenter R2 SP2 64-bit

Windows Server 2003 DataCenter SP1 32-bit

Windows Server 2003 DataCenter SP2 32-bit

Windows Server 2003 Enterprise SP1 32-bit

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 101

Type OS

Windows Server 2003 Enterprise SP2 32/64-bit

Windows Server 2003 Standard SP1 32-bit

Windows Server 2003 Standard SP2 32/64-bit

Linux Red Hat Enterprise Linux 5.4/5.5 64-bit

Red Hat Enterprise Linux 6.1/6.2/6.4/6.5 64-bit

A.2 Parameter Reference

A.2.1 Cluster Parameters

Table A-3 Parameter description

Parameter Description Modifiable Example

Value

Cluster name Specifies the name of the cluster to be Yes cluster01

created.
This parameter is mandatory.

Description Provides supplementary information Yes N/A

about the cluster to be created.
This parameter is optional.

Host memory Specifies the host memory Yes N/A

overcommit overcommitment function. If this function
ment is enabled, the memory provided by the
host for VMs can be greater than the
physical memory of the host, and
therefore the VM density supported by
the host can be increased.
This parameter is optional.

VM startup Specifies the VM startup policy. Yes Assign based

policy Value: on load
balancing
l Assign automatically: The system
starts a VM on a host that has
available resources in the cluster.
l Assign based on load balancing: The
system starts VMs on the host with the
largest available CPU capacity.
This parameter is optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 102

Parameter Description Modifiable Example

Value

GuestNUMA Guest NUMA presents a topology view Yes N/A

of memory and CPU resources on each
host to VMs, the VM user can configure
VM CPUs and memory using the third-
party software (such as Eclipse) based on
this topology, so that VMs can obtain the
most easy-to-access memory based on the
topology, thereby reducing access latency
and improving VM performance.
This parameter is optional.

VMs Specifies the policy of processing the VM Yes No processing

processing if the data store used by the VM is faulty.
policy upon Value:
data store
faults l No processing
l Stop VM
This parameter is optional.

Local Specifies whether hosts in the cluster can Yes N/A

Ramdisk use local RAM disks.
Values:
l Enable
l Disable
This parameter is optional.

Configure Specifies whether to enable the high Yes N/A

HA availability (HA) function.
Value:
l HA resource reservation
l Tolerate cluster host failures
This parameter is optional.

Start resource Specifies whether to enable the resource Yes N/A

scheduling scheduling function, with which you can
set the automation level and migration
threshold.
This parameter is deselected by default.
This parameter is optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 103

Parameter Description Modifiable Example

Value

Automation Specifies the automation level for Yes Automatic

level resource scheduling.
This parameter is valid only when
Enable computing resource scheduling
is selected.
Value:
l Manual: The system provides VM
migration suggestions based on the
resource load, and users can determine
whether to migrate VMs.
l Automatic: The system automatically
migrates VMs to maximize resource
utilization when the resource load
exceeds the migration threshold.
This parameter is optional.

Measured by Specifies the resource type by which Yes CPU

resource scheduling is performed.
This parameter is valid only when Start
resource scheduling is selected.
Value:
l CPU
l Memory
l CPU and Memory
This parameter is optional.

Migration Specifies the migration threshold for Yes Medium

Threshold resource scheduling.
This parameter is valid only when Start
resource scheduling is selected.
Value:
l Conservative: The system does not
migrate VMs or provide VM
migration suggestions.
l Slightly conservative, Medium,
Slightly radical, and Radical: The
system improves cluster load
imbalance by a degree in ascending
order.
Default value: Medium
This parameter is optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 104

Parameter Description Modifiable Example

Value

Resource Specifies the period during which the Yes By day

scheduling specified thresholds take effect.
interval Value:
l By day
l By week
l By month
This parameter is optional.

VM Specifies the VM scheduling automation Yes N/A

scheduling level. Each VM in a cluster can have an
customizatio individual automation level configured.
n This parameter is optional.

Enable Specifies whether to enable the automatic Yes N/A

automated power management function for a host.
power With this function enabled, the system
management migrates the VMs on the host and powers
on or off the host based on the resource
usage on the host.
This function is remotely implemented by
the Baseboard Management Controller
(BMC). Therefore, you need to configure
the BMC for the host before enabling this
function on it.
This parameter is optional.

Automation Specifies the automation level for Yes Automatic

level resource scheduling. This parameter is
valid only when Enable automated
power management is selected.
Value:
l Manual: The system provides VM
migration suggestions based on the
resource load, and users can determine
whether to migrate VMs.
l Automatic: The system automatically
migrates VMs to maximize resource
utilization when the resource load
exceeds the migration threshold.
This parameter is optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 105

Parameter Description Modifiable Example

Value

Power Specifies the power management Yes Medium

Management thresholds. This parameter is valid only
Threshold when Enable automated power
management is selected.
Value:
l Conservative: The system does not
power off hosts by default. It powers
on available hosts in the cluster only
when the average host resource usage
in the cluster is higher than the heavy-
load threshold.
l Slightly conservative, Medium and
Slightly radical: The system powers
on available hosts in the cluster when
the average host resource usage in the
cluster is higher than the heavy-load
threshold. It powers off some hosts
when the average host resource usage
in the cluster is lower than the light-
load threshold.
l Radical: The system does not power
on available hosts by default. It
powers off some hosts in the cluster
only when the average host resource
usage in the cluster is lower than the
light-load threshold.
This parameter is optional.

Resource Specifies the period during which the Yes By day

scheduling specified thresholds take effect.
interval Value:
l By day
l By week
l By month
This parameter is optional.

Enable IMC Specifies the incompatible migration Yes N/A

cluster (IMC) function. The IMC function
allows the VMs running on the hosts in
the cluster to present the same CPU
features. This ensures successful VM
migration across these hosts even if the
hosts use physical CPUs of different
performance baselines.
This parameter is optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 106

A.2.2 Computing Resource Scheduling Rules Parameters

Table A-4 Parameter description
Category Parameter Description Modifiab Exa
le mpl
e
Val
ue

Basic Enable Specifies whether to enable computing Yes N/A

informatio computing resources scheduling.
n resource Other parameters are available and the
scheduling resource scheduling policies can be
configured only after this parameter is
selected.
This parameter is optional.

Automation Specifies the automation level for VM Yes Aut

level migration. oma
This parameter is valid only when Enable tic
computing resource scheduling is
selected.
Value:
l Manual: The system provides resource
scheduling suggestions to implement
load balancing of the cluster.
l Automatic: The system automatically
migrates VMs to maximize resource
utilization.
Default value: Automatic
This parameter is optional.

Measured Specifies the resource type by which Yes CPU

by resource scheduling is performed. The &
computing resource usage includes the Me
usage of the CPU and memory resources mor
in the cluster. y
This parameter is valid only when Enable
computing resource scheduling is
selected.
Value:
l CPU
l Memory
l CPU & Memory
Default value: CPU & Memory
This parameter is optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 107

Category Parameter Description Modifiab Exa

le mpl
e
Val
ue

Migration Specifies a value that determines whether Yes Med

Threshold the system migrates VMs or provides ium
migration suggestions.
This parameter is valid only when Enable
computing resource scheduling is
selected.
Each column in the table indicates a
period of one hour. You can set a
migration threshold for each period.
Value:
l Conservative: The system does not
migrate VMs or provide VM
migration suggestions.
l Slightly conservative, Medium,
Slightly radical, and Radical: The
system improves cluster load
imbalance by a degree in ascending
order.
This parameter is optional.

Resource Specifies the period during which the Yes By

scheduling specified thresholds take effect. day
interval Value:
l By day
l By week
l By month
This parameter is optional.

VM Allows users to set a scheduling Yes N/A

scheduling automation level for each VM in a cluster.
customizati This parameter is optional.
on

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 108

Category Parameter Description Modifiab Exa

le mpl
e
Val
ue

Enable Specifies whether to enable the Yes N/A

automated automated power management function.
power With this function enabled, the system
manageme migrates VMs on the host based on the
nt host resource usage, and powers off or on
the hosts in the cluster as needed.
The automated power management
function uses the host BMC to remotely
manage host power settings. Therefore,
before enabling this function, you must
configure the BMC for each host.
This parameter is optional.

Automation Specifies the scheduling mode. This Yes Aut

level parameter is valid only when Enable oma
computing resource scheduling is tic
selected.
Value:
l Manual: The system provides VM
migration suggestions based on the
resource load, and users can determine
whether to migrate VMs.
l Automatic: The system automatically
migrates VMs to maximize resource
utilization when the resource load
exceeds the migration threshold.
This parameter is optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 109

Category Parameter Description Modifiab Exa

le mpl
e
Val
ue

Power Specifies the power management Yes Med

Manageme thresholds. This parameter is valid only ium
nt when Enable automated power
Threshold management is selected.
Value:
l Conservative: The system does not
power off hosts by default. It powers
on available hosts in the cluster only
when the average host resource usage
in the cluster is higher than the heavy-
load threshold.
l Slightly conservative, Medium and
Slightly radical: The system powers
on available hosts in the cluster when
the average host resource usage in the
cluster is higher than the heavy-load
threshold. It powers off some hosts
when the average host resource usage
in the cluster is lower than the light-
load threshold.
l Radical: The system does not power
on available hosts by default. It
powers off some hosts in the cluster
only when the average host resource
usage in the cluster is lower than the
light-load threshold.
This parameter is optional.

Resource Specifies the period during which the Yes By

scheduling specified thresholds take effect. day
interval Value:
l By day
l By week
l By month
This parameter is optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 110

Category Parameter Description Modifiab Exa

le mpl
e
Val
ue

Resource VM Group Specifies a VM group in the cluster. Yes VM

group This parameter is valid only when Enable grou
computing resource scheduling is p01
selected.
You can click Add under VM group to
create a VM group for VMs in the cluster,
so that the VM group rules can be created
or modified.
This parameter is optional.

Host Group Specifies a host group in the cluster. Yes host

This parameter is valid only when Enable grou
computing resource scheduling is p01
selected.
You can click Add under Host group to
create a host group for hosts in the
cluster.
This parameter is mandatory.

Rule group Name Specifies the name of the rule for VM Yes rule
migration. 01
This parameter is valid only when Enable
computing resource scheduling is
selected.
This parameter is optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 111

Category Parameter Description Modifiab Exa

le mpl
e
Val
ue

Type Specifies the type of the rule for VM Yes Mut

migration. uall
This parameter is valid only when Enable y
computing resource scheduling is excl
selected. usiv
e
Value: VM
l Keep VMs together: This rule keeps s
the selected VMs always running on
the same host. One VM can be added
to only one keep-VMs-together rule.
l Mutually exclusive VMs: This rule
keeps the selected VMs running on
different hosts. One VM can be added
to only one VM-mutually-exclusive
rule.
l VMs to hosts: This rule associates a
VM group with a host group so that
the specified VMs in the VM group
can be configured to run only on the
specified host in the host group.
This parameter is optional.

Balancing Name Specifies the name of a load balancing Yes N/A

Group group.
This parameter is optional.

Advanced Setting Specifies the name of the advanced Yes Max

Settings setting option. For example, the Mig
MaxMigrationNumber option can be ratio
used to set the maximum number of VMs nNu
that can be migrated at a time. The mbe
default number is 40. r
For information about other advanced
setting options and values, contact
technical support.
This parameter is optional.

Value Specifies the value of an advanced setting Yes 30

option.
This parameter is optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 112

A.2.3 HA Parameters
Table A-5 Parameter description
Parameter Description Modifiable Example Value

HA The high availability Yes N/A

(HA) function can
be enabled for VMs
in a cluster only
when HA is enabled
for the cluster.
This parameter is
optional.

HA resource If this function is Yes N/A

reservation enabled for a cluster,
the system reserves
preset CPU and
memory resources in
the cluster to ensure
VM high
availability.
This parameter is
optional.

Reserved CPU (%) Specifies the Yes 5

percentage of the
CPU resources
reserved to the total
CPU resources in
the cluster.
This parameter is
optional.

Reserved memory Specifies the Yes 5

(%) percentage of the
memory resources
reserved to the total
memory resources in
the cluster.
This parameter is
optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 113

Use dedicated The system reserves Yes N/A

failover hosts some hosts as
dedicated failover
hosts. VMs on a
failover host cannot
be started, migrated,
woken up, or
restored using a
snapshot.
If an HA-enabled
VM fails, the system
restarts the VM on a
common host or a
failover host based
on resource usage on
the hosts.
If Migrate all VMs
on the dedicated
failover host is
selected, the system
periodically
migrates VMs on a
failover host to other
common hosts that
have sufficient
resources to reserve
resources on the
failover host.
This parameter is
optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 114

Tolerate cluster host Specifies the Yes N/A

failures number of tolerated
faulty hosts in a
cluster. The system
periodically checks
whether resources in
the cluster are
sufficient for VMs
failover on these
hosts. If the
resources are
insufficient, the
system generates an
alarm to notify the
user.
Values:
l Automatic: The
system selects
the maximum
reserved CPU
resources (MHz)
and memory
resources
(MB/GB) on all
running VMs in
the cluster as the
CPU slot size
and the memory
slot size.
l Custom: You
need to set a
custom CPU slot
size and a
memory slot size
based on site
requirements.
This parameter is
optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 115

A.2.4 VM Startup Policy Parameters

Table A-6 Parameter description
Parameter Description Modifiable Example Value

VM startup policy Values: Yes Assign based on

l Assign load balancing
automatically:
The system starts
a VM on a host
that has available
resources in the
cluster.
l Assign based on
load balancing:
The system starts
a VM on a host
that has lower
CPU and
memory usage.
When computing
resource scheduling
is enabled for the
cluster, this
parameter is
unavailable, and the
system uses the
default VM startup
policy Assign based
on load balancing.
This parameter is
optional.

A.2.5 Host Parameters

Table A-7 Parameter description
Catego Paramete Description Modifia Example
ry r ble Value

Host Name Specifies the name of the host. Yes CNA01

This parameter is mandatory.

IP address Specifies the management IP address of Yes 192.168.1

the host. 0.1
This parameter is mandatory.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 116

Catego Paramete Description Modifia Example

ry r ble Value

Descriptio Provides supplementary information Yes N/A

n about the host.
This parameter is optional.

BMC BMC IP Specifies the baseboard management Yes 192.168.1

controller (BMC) IP address of the host. 1.1
This parameter is optional.

Username Specifies the username for logging in to Yes root

the BMC system of the host.
This parameter is optional.

Password Specifies the password for logging in to Yes Huawei1

the BMC system of the host. 2#$
This parameter is optional.

A.2.6 Time Synchronization Parameters

Table A-8 Parameter description
Parameter Description Modifiable Example Value

NTP server Specifies the IP Yes 192.168.40.3

address or domain
name of the NTP
server.
A minimum of one
or a maximum of
three NTP servers
can be configured.
This parameter is
mandatory.

Synchronization Specifies the time Yes 64

interval(s) synchronization
interval.
This parameter is
optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 117

A.2.7 BMC Parameters

Table A-9 Parameter description
Parameter Description Modifiable Example Value

IP address Specifies the BMC Yes 192.168.200.20

IP address of the
host.
This parameter is
optional.

Username Specifies the BMC Yes root

username of the
host.
This parameter is
optional.

Password Specifies the BMC Yes Huawei12#$

password of the
host.
This parameter is
optional.

A.2.8 Network Ports Parameters

Table A-10 Parameter description
Param Description Modifiable Example
eter Value

Name Specifies the name of the Yes band001

network ports to be bound to
the host.
This parameter is mandatory.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 118

Param Description Modifiable Example

eter Value

Bindin Specifies the binding mode. Yes Active/Standby

g The following binding modes
Mode are available for common
network interface cards
(NICs):
l Active-backup: applies to
scenarios where two
network ports are to be
bound. This mode
provides high reliability.
The bandwidth of the
bound port in this mode
equals to that of a member
port.
l Round-robin: applies to
scenarios where two or
more network ports are to
be bound. The bandwidth
of the bound port in this
mode is higher than that
of a member port, because
the member ports share
workloads in sequence.
This mode may result in
data packet disorder
because traffic is evenly
sent to each port.
Therefore, MAC address
based load balancing
prevails over Polling in
load sharing modes.
l IP address and port-
based load balancing:
applies to scenarios where
two or more network ports
are to be bound. The
bandwidth of the bound
port in this mode is higher
than that of a member
port, because the member
ports share workloads
based on the source-
destination-port-based
load sharing algorithm.
Source-destination-port-
based load balancing
algorithm: When the

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 119

Param Description Modifiable Example

eter Value

packets contain IP
addresses and ports, the
member ports share loads
based on the source and
destination IP addresses,
ports, and MAC
addresses. When the
packets contain IP
addresses, the member
ports share loads based on
the IP addresses and MAC
addresses. When the
packets contain only
MAC addresses, the
member ports share loads
based on the MAC
addresses.
This mode is
recommended when the
virtual extensible LAN
(VXLAN) function is
enabled. This mode
allows network traffic to
be evenly distributed
based on the source and
destination port
information in the
packets.
l MAC address-based
load balancing: applies to
scenarios where two or
more network ports are to
be bound. The bandwidth
of the bound port in this
mode is higher than that
of a member port, because
the member ports share
workloads based on the
MAC addresses of the
source and destination
ports.
This mode is
recommended when most
network traffic is on the
layer 2 network. This
mode allows network
traffic to be evenly

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 120

Param Description Modifiable Example

eter Value

distributed based on the

MAC addresses.
l MAC address-based
LACP: This mode is
developed based on the
MAC address based
load balancing mode. In
MAC address-based
LACP mode, the bound
port can automatically
detect faults on the link
layer and trigger a
switchover if a link fails
using the LACP protocol.
l IP address-based LACP:
applies to scenarios where
two or more network ports
are to be bound. The
bandwidth of the bound
port in this mode is higher
than that of a member
port, because the member
ports share workloads
based on the source-
destination-IP-address-
based load sharing
algorithm. When the
packets contain IP
addresses, the member
ports share loads based on
the IP addresses and MAC
addresses. When the
packets contain only
MAC addresses, the
member ports share loads
based on the MAC
addresses. In this mode,
the bound port can also
automatically detect faults
on the link layer and
trigger a switchover if a
link fails using the LACP
protocol.
This mode is
recommended when most
network traffic goes
across layer 2 and layer 3
networks.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 121

Param Description Modifiable Example

eter Value

The following binding modes

are available for intelligent
network interface cards
(iNICs):
l Active-backup: applies to
scenarios where two
network ports are to be
bound. This mode
provides high reliability.
The bandwidth of the
bound port in this mode
equals to that of a member
port.
l Source MAC address-
based load balancing:
applies to scenarios where
two or more network ports
are to be bound. The
bandwidth of the bound
port in this mode is higher
than that of a member
port, because the member
ports share workloads
based on the MAC
address of the source port.
l Destination MAC
address-based load
balancing: applies to
scenarios where two or
more network ports are to
be bound. The bandwidth
of the bound port in this
mode is higher than that
of a member port, because
the member ports share
workloads based on the
MAC address of the
destination port.
This mode is
recommended when most
network traffic is on the
layer 2 network. This
mode allows network
traffic to be evenly
distributed based on the
MAC addresses.
l Source IP address-based
load balancing: applies to

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 122

Param Description Modifiable Example

eter Value

scenarios where two or

more network ports are to
be bound. The bandwidth
of the bound port in this
mode is higher than that
of a member port, because
the member ports share
workloads based on the IP
address of the source port.
l Destination IP address-
based load balancing:
applies to scenarios where
two or more network ports
are to be bound. The
bandwidth of the bound
port in this mode is higher
than that of a member
port, because the member
ports share workloads
based on the IP address of
the destination port.
This mode is
recommended when most
network traffic is on the
layer 3 network. This
mode allows network
traffic to be evenly
distributed based on the
destination IP addresses.
This parameter is mandatory.

A.2.9 Storage Port Parameters

Table A-11 Parameter description

Param Description Modifiable Example

eter Value

Name Specifies the name of the Yes iscsi0

storage port to be added.
This parameter is mandatory.

Descri Provides supplementary Yes N/A

ption information about the storage
port.
This parameter is optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 123

Param Description Modifiable Example

eter Value

IP Specifies the IP address of Yes 172.20.100.200

addres the storage port. It must be an
s idle IP address in the same
network segment as the
storage IP address of the
storage device.
This parameter is mandatory.

Subnet Specifies the subnet mask of Yes 255.255.0.0

mask the storage plane.
This parameter is mandatory.

VLAN Specifies the virtual local Yes 4070

ID area network (VLAN) ID of
the storage plane.
This parameter is optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 124

Param Description Modifiable Example

eter Value

Outbo Specifies the limit for the Yes -

und send traffic on the system
Traffic port.
Shapin Value range:
g
l Average send bandwidth
(Mbit/s): 1 to 10000
l Peak send bandwidth
(Mbit/s): Average send
bandwidth to 10000
l Burst send size (Mbit/s):
Peak send bandwidth to
10000
This parameter is optional.

A.2.10 System Port Parameters

Table A-12 Parameter description
Paramet Description Modifiable Exa
er mple
Valu
e

Name Specifies the name of the system port. Yes Mana

This parameter is optional. geme
nt

Descripti Provides supplementary information about the Yes Mana

on system port. geme
This parameter is optional. nt
port

IP Specifies the IP address of the system port. Yes 192.1

address This parameter is optional. 68.20
0.10

Subnet Specifies the subnet mask of the system port. Yes 255.2
mask This parameter is optional. 55.25
5.0

Gateway Specifies the gateway of a management port. Yes 192.1

This parameter is optional. 68.20
0.1

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 125

Paramet Description Modifiable Exa

er mple
Valu
e

Router Specifies the router information, including: Yes Netw

l Network destination: specifies the first IP ork
address of the network segment that the desti
target storage devices use, for example, natio
192.168.0.0. n:
192.1
l Netmask: specifies the subnet mask of the 68.0.
network segment that the target storage 0
devices use.
Netm
l Gateway: specifies the gateway of the ask:
storage ports. The gateway address must be 255.2
set on the network segment providing storage 55.25
IP addresses. 5.0
The router information only applies to storage Gate
ports. way:
This parameter is mandatory when the storage 172.2
plane requires layer 3 communications. 0.30.
1

VLAN Identifies the VLAN of a system port. No 4070

ID Set it to the VLAN ID of the management plane
for a management port and to the VLAN ID of
the storage plane for a storage port.
This parameter is optional.

Outboun Specifies the limit for the send traffic on the Yes -
d Traffic system port.
Shaping Value range:
l Average send bandwidth (Mbit/s): 1 to 10000
l Peak send bandwidth (Mbit/s): Average send
bandwidth to 10000
l Burst send size (Mbit/s): Peak send
bandwidth to 10000
This parameter is optional.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 126

A.3 Parameters in Advanced Settings for Computing

Resource Scheduling
Parameter Description Attribute Value Range Default Value

MaxMigration Specifies the int (0, +∞) 40

Number number of
maximum VMs
to be migrated
in a batch. This
parameter does
not apply to
VMs that are
migrated based
on the
computing
resource
scheduling rules
(such as the
Mutually
Exclusive rule
or Keep VMs
Together rule).

EnableCpuRese Specifies boolean TRUE or TRUE

rve whether to FALSE (indicating
enable the CPU enabled)
reservation
function. If the
function is
enabled, the
distributed
resource
scheduling
(DRS) module
calculates both
the actual
workloads and
CPU
reservation rate
to provide VM
migration
suggestions.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 127

Parameter Description Attribute Value Range Default Value

EnableMemory Specifies boolean TRUE or TRUE

Reserve whether to FALSE (indicating
enable the enabled)
memory
reservation
function. If the
function is
enabled, the
DRS module
calculates both
the actual
workloads and
memory
reservation rate
to provide VM
migration
suggestions.

MigrationThres Specifies the int [1,2] 1

holdsType elements based
on which the
migration
threshold is
determined,
including:
l Host loads.
l Host loads
and reserved
resources.

PowerOnEvalu Specifies the int [60, 3600] 300

ationTime period of time
(in seconds)
used to evaluate
whether to
power on hosts.

PowerOffHostE Specifies the int [60, 3600] 2400

valuationTime period of time
(in seconds)
used to evaluate
whether to
power off hosts.

TinyVmSize Specifies the int [0, 40960] 512

memory size
(MB) of a tiny
VM.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 128

Parameter Description Attribute Value Range Default Value

TinyVmMigrati Specifies the int [0, 3600] 30

onTime migration
duration (in
minutes) of a
tiny VM.

SmallVmSize Specifies the int [0, 40960] 1024

memory size
(MB) of a
small-sized
VM.

SmallVmMigra Specifies the int [0, 3600] 60

tionTime migration
duration (in
minutes) of a
small-sized
VM.

MiddleVmSize Specifies the int [0, 40960] 2048

memory size
(MB) of a
medium-sized
VM.

MiddleVmMigr Specifies the int [0, 3600] 90

ationTime migration
duration (in
minutes) of a
medium-sized
VM.

LargeVmSize Specifies the int [0, 40960] 4096

memory size
(MB) of a
large-sized VM.

LargeVmMigrat Specifies the int [0, 3600] 120

ionTime migration
duration (in
minutes) of a
large-sized VM.

ReservedUnhig Specifies the int [0, 128] 1

hLoadHostNum number of
ber reserved non-
light-load hosts
in a cluster.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 129

Parameter Description Attribute Value Range Default Value

LoadBalanceVa Specifies the float [0.0, 1.0] 0.0005

lueDecreaseThr threshold for
eshold the load
balancing
function. The
system
generates a VM
migration
suggestion for
load balancing
only when the
load balancing
level in the
cluster drops
below this
threshold.

LoadDataFTTh Specifies the int [0, 100] 20

reshold fault tolerance
threshold (%)
for load data.
If the load data
collection
failure rate
exceeds the
threshold within
the specified
fault tolerance
interval,
scheduling
fails.

LoadDataFTInt Specifies the int [60, 3600] 600

erval fault tolerance
interval (in
seconds) the
load data.
If the load data
collection
failure rate
exceeds the
fault tolerance
threshold within
the specified
interval,
scheduling
fails.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 130

NOTE

FusionCompute defines five types of VM memory specifications (tiny VM, small-sized VM, medium-
sized VM, and large-sized VM) and migration duration. The DRS function calculates the VM migration
costs based on the VM memory specification and migration duration and does not migrate VMs with
high migration costs.
If a cluster accommodates VMs with different memory specifications and migration duration, you can
change the value ranges of the VM memory specifications and migration duration.

Issue 01 (2015-11-11) Huawei Proprietary and Confidential 131

Moho Users Manual - HQ
No ratings yet
Moho Users Manual - HQ
597 pages
U2020 Virtual Cluster System Administrator Guide (EulerOS, TaiShan) (V300R019C10 - 07) (PDF) - en
No ratings yet
U2020 Virtual Cluster System Administrator Guide (EulerOS, TaiShan) (V300R019C10 - 07) (PDF) - en
628 pages
FusionCloud Desktop V100R005C20 Solution Description 03
No ratings yet
FusionCloud Desktop V100R005C20 Solution Description 03
136 pages
16 FusionSphare OpenStack OM Services Operation
No ratings yet
16 FusionSphare OpenStack OM Services Operation
37 pages
NetEngine 8000 M14, M8 and M4 V800R023C00SPC500 Configuration Guide 02 System Management
No ratings yet
NetEngine 8000 M14, M8 and M4 V800R023C00SPC500 Configuration Guide 02 System Management
825 pages
FusionModule800 Smart Small Data Center V100R001C20 User Manual (NetCol5000-A013)
No ratings yet
FusionModule800 Smart Small Data Center V100R001C20 User Manual (NetCol5000-A013)
629 pages
Huawei SAN Storage Host Connectivity Guide For Windows
No ratings yet
Huawei SAN Storage Host Connectivity Guide For Windows
120 pages
08 Huawei Virtualization Platform Management and Usage
No ratings yet
08 Huawei Virtualization Platform Management and Usage
70 pages
FusionCompute V100R005C00 Installation and Configuration PDF
No ratings yet
FusionCompute V100R005C00 Installation and Configuration PDF
39 pages
FusionCompute V100R005C10 Network Management Guide 01
No ratings yet
FusionCompute V100R005C10 Network Management Guide 01
82 pages
SUSE Linux Cluster
No ratings yet
SUSE Linux Cluster
392 pages
FusionSphere PCS 6.5.1.1 Operation Guide (HCS Xen To KVM) 01
No ratings yet
FusionSphere PCS 6.5.1.1 Operation Guide (HCS Xen To KVM) 01
142 pages
HCIA Cloud Computing V4.0 Lab Guide PDF
No ratings yet
HCIA Cloud Computing V4.0 Lab Guide PDF
133 pages
Adobe InDesign 2.0
No ratings yet
Adobe InDesign 2.0
444 pages
Compuverde Management Manual 1.7.1.1
No ratings yet
Compuverde Management Manual 1.7.1.1
123 pages
07 Lab Guide-Student
No ratings yet
07 Lab Guide-Student
45 pages
FusionSphere V100R005C10U1 Software Installation Guide (Server Virtualization) 01
No ratings yet
FusionSphere V100R005C10U1 Software Installation Guide (Server Virtualization) 01
150 pages
NE9000 V800R023C00SPC500 Configuration Guide 02 System Management
No ratings yet
NE9000 V800R023C00SPC500 Configuration Guide 02 System Management
607 pages
CL15-01 FusionSphere Overview
No ratings yet
CL15-01 FusionSphere Overview
30 pages
Useful FusionCompute V100R005C10 Initial Configuration Guide 01
No ratings yet
Useful FusionCompute V100R005C10 Initial Configuration Guide 01
132 pages
FusionCompute V100R005C10 Quick Start 01
No ratings yet
FusionCompute V100R005C10 Quick Start 01
48 pages
Introduction To Virtualization Features
No ratings yet
Introduction To Virtualization Features
31 pages
HCNA-Cloud V2.0 Lab Guide
No ratings yet
HCNA-Cloud V2.0 Lab Guide
181 pages
FusionServer Pro Rack Server IBMC (V250 To V259) User Guide 31
No ratings yet
FusionServer Pro Rack Server IBMC (V250 To V259) User Guide 31
499 pages
VPSS3i ARBEND Users Guide
No ratings yet
VPSS3i ARBEND Users Guide
183 pages
EcoStruxure Building Operation - IT Reference Guide
No ratings yet
EcoStruxure Building Operation - IT Reference Guide
602 pages
Fortigate Ha 524
No ratings yet
Fortigate Ha 524
302 pages
FusionServer Pro Rack Server iBMC (V290 To V299) User Guide 27
No ratings yet
FusionServer Pro Rack Server iBMC (V290 To V299) User Guide 27
525 pages
Huawei SAN Storage
No ratings yet
Huawei SAN Storage
153 pages
FusionSphere V100R005C00 Quick User Guide 01
No ratings yet
FusionSphere V100R005C00 Quick User Guide 01
64 pages
FusionModule800 Smart Small Data Center V100R021C00 User Manual
No ratings yet
FusionModule800 Smart Small Data Center V100R021C00 User Manual
508 pages
07 Huawei Virtualization Platform
No ratings yet
07 Huawei Virtualization Platform
29 pages
Differentiate Between Graphical User Interface and Web Page Design
0% (1)
Differentiate Between Graphical User Interface and Web Page Design
4 pages
02 Compute Services
No ratings yet
02 Compute Services
30 pages
02 InstallAndConfig EG
No ratings yet
02 InstallAndConfig EG
5 pages
Book-Sleha-Guide Color en PDF
No ratings yet
Book-Sleha-Guide Color en PDF
369 pages
Book-Sleha Color en PDF
No ratings yet
Book-Sleha Color en PDF
394 pages
XXXXX: (Autonomous)
100% (1)
XXXXX: (Autonomous)
26 pages
Imaster NCE-Campus V300R020C10 Monitoring and O&M
No ratings yet
Imaster NCE-Campus V300R020C10 Monitoring and O&M
96 pages
FusionModule2000 Smart Modular Data Center V500R003C10 Installation Guide
No ratings yet
FusionModule2000 Smart Modular Data Center V500R003C10 Installation Guide
338 pages
Book Sleha Guide Color en
No ratings yet
Book Sleha Guide Color en
368 pages
FusionCompute V100R005C10 Storage Management Guide 01
No ratings yet
FusionCompute V100R005C10 Storage Management Guide 01
154 pages
Configuring The Controller Documentation.
No ratings yet
Configuring The Controller Documentation.
8 pages
Fusion Module 500
No ratings yet
Fusion Module 500
257 pages
FusionSphere V100R003C00 Quick Installation and Configuration Guide 01 PDF
100% (1)
FusionSphere V100R003C00 Quick Installation and Configuration Guide 01 PDF
33 pages
IYC05 FusionCompute Routine Maintenance
No ratings yet
IYC05 FusionCompute Routine Maintenance
36 pages
HUAWEI SAN Storage Host Connectivity Guide For Windows
No ratings yet
HUAWEI SAN Storage Host Connectivity Guide For Windows
88 pages
01 HCNA-Cloud V2.0 Lab Guide
No ratings yet
01 HCNA-Cloud V2.0 Lab Guide
179 pages
Ontap 8 Sysadmin
No ratings yet
Ontap 8 Sysadmin
192 pages
FusionServer Tools V2R2 ServiceCD 2.0 User Guide 20
No ratings yet
FusionServer Tools V2R2 ServiceCD 2.0 User Guide 20
220 pages
I Master
No ratings yet
I Master
29 pages
FusionDC1000A40 Prefabricated All-in-One Data Center Quick Guide (2021, Zambia, ZESCO)
No ratings yet
FusionDC1000A40 Prefabricated All-in-One Data Center Quick Guide (2021, Zambia, ZESCO)
30 pages
HFSM and HDI Cluster Initial Configuration - v1-0
No ratings yet
HFSM and HDI Cluster Initial Configuration - v1-0
79 pages
OS MCQ Combine
No ratings yet
OS MCQ Combine
30 pages
FusionCompute V100R005C10 Product Description 01
No ratings yet
FusionCompute V100R005C10 Product Description 01
61 pages
Advanced Setup Guide AOS v5 - 19
No ratings yet
Advanced Setup Guide AOS v5 - 19
28 pages
Question Bank: Mangalore University Computer Science Teachers' Association
No ratings yet
Question Bank: Mangalore University Computer Science Teachers' Association
8 pages
OpenShift - Container - Platform 4.5 Nodes en US PDF
No ratings yet
OpenShift - Container - Platform 4.5 Nodes en US PDF
248 pages
FusionSphere6.5.0 Installation and Configuration
No ratings yet
FusionSphere6.5.0 Installation and Configuration
41 pages
02 Introduction To Compute Virtualization
No ratings yet
02 Introduction To Compute Virtualization
30 pages
FusionSphere V100R003C10 Quick User Guide (Standalone Mode) 02
No ratings yet
FusionSphere V100R003C10 Quick User Guide (Standalone Mode) 02
38 pages
Hyper Historian Getting Started Guide
No ratings yet
Hyper Historian Getting Started Guide
117 pages
Huawei FusionSphere 5.1 Brochure (Server Virtualization)
No ratings yet
Huawei FusionSphere 5.1 Brochure (Server Virtualization)
2 pages
CL02 FusionCompute Product Introduction
No ratings yet
CL02 FusionCompute Product Introduction
41 pages
Unity in Action Multiplatform Game Development in C With Unity 5 1st Edition Joseph Hocking Instant Download
No ratings yet
Unity in Action Multiplatform Game Development in C With Unity 5 1st Edition Joseph Hocking Instant Download
52 pages
Grade 4 Second Quarter Lecture 2 2020
No ratings yet
Grade 4 Second Quarter Lecture 2 2020
7 pages
AS 82545 KVCOM Forexcel UM 96119E GB WW 1086-9
No ratings yet
AS 82545 KVCOM Forexcel UM 96119E GB WW 1086-9
272 pages
(IJCST-V12I1P5) :debashis Dey, Samiran Das, Pallavi Thakuri, Dr. Bornali Gogoi, Prof. Nelson Varte
No ratings yet
(IJCST-V12I1P5) :debashis Dey, Samiran Das, Pallavi Thakuri, Dr. Bornali Gogoi, Prof. Nelson Varte
6 pages
SAP HANA Cockpit Installation Guide en
No ratings yet
SAP HANA Cockpit Installation Guide en
54 pages
Pysimplegui 2020 Clearly Seeing A Gui For You: The Basic Questions: Who
No ratings yet
Pysimplegui 2020 Clearly Seeing A Gui For You: The Basic Questions: Who
42 pages
Psychology Thesis Defense by Slidesgo
No ratings yet
Psychology Thesis Defense by Slidesgo
45 pages
ANI1
No ratings yet
ANI1
4 pages
THE Information AGE
No ratings yet
THE Information AGE
79 pages
Kernel Mode Vs User Mode
No ratings yet
Kernel Mode Vs User Mode
29 pages
Speed Up Ubuntu PDF
No ratings yet
Speed Up Ubuntu PDF
5 pages
HedyMed综合画册
No ratings yet
HedyMed综合画册
8 pages
Full Pivotcall Options Course Available Premiumcourses12: Generated Via PDF Scanner
No ratings yet
Full Pivotcall Options Course Available Premiumcourses12: Generated Via PDF Scanner
34 pages
UNIT 4 Introduction To Linux & Shell Programming
No ratings yet
UNIT 4 Introduction To Linux & Shell Programming
13 pages
Oceanspace S12000
No ratings yet
Oceanspace S12000
2 pages
Final Autocad Manual Made by Warid Javed
No ratings yet
Final Autocad Manual Made by Warid Javed
5 pages
A Review of Visual SLAM For Robotics - Evolution, Properties, and Future Applications - Frobt-11-1347985
No ratings yet
A Review of Visual SLAM For Robotics - Evolution, Properties, and Future Applications - Frobt-11-1347985
18 pages
Bachelor of Computer Application (BCA) Bca/Assign/Iii/Year/ Assignments
No ratings yet
Bachelor of Computer Application (BCA) Bca/Assign/Iii/Year/ Assignments
7 pages
JAVA Mini Project
No ratings yet
JAVA Mini Project
6 pages
BizCardX - Extracting Business Card Data With OCR
No ratings yet
BizCardX - Extracting Business Card Data With OCR
3 pages
CISP Appa
No ratings yet
CISP Appa
4 pages
Mastering Linux Security and Hardening - Second Edition: Protect your Linux systems from intruders, malware attacks, and other cyber threats, 2nd Edition
From Everand
Mastering Linux Security and Hardening - Second Edition: Protect your Linux systems from intruders, malware attacks, and other cyber threats, 2nd Edition
Donald A. Tevault
No ratings yet
CompTIA Security+ SY0-601 Exam Practice Tests
From Everand
CompTIA Security+ SY0-601 Exam Practice Tests
CertSquad Professional Trainers
No ratings yet
IT Survival Guide !!!: Avoid Disaster, Cut Cost, Protect Your Sh*t!
From Everand
IT Survival Guide !!!: Avoid Disaster, Cut Cost, Protect Your Sh*t!
Alp Ertunga Isin
No ratings yet
Implementing VMware Dynamic Environment Manager: Manage, Administer and Control VMware DEM, Dynamic Desktop, User Policies and Complete Troubleshooting
From Everand
Implementing VMware Dynamic Environment Manager: Manage, Administer and Control VMware DEM, Dynamic Desktop, User Policies and Complete Troubleshooting
Peter von Oven
No ratings yet
Instant Getting Started with VMware Fusion
From Everand
Instant Getting Started with VMware Fusion
Michael Roy
No ratings yet