VPLEX SolVe Generator

Solution for Validating your engagement

Topic
VPLEX Installation and Upgrade Procedures
Selections
Upgrade Procedures: Upgrade GeoSynchrony
Upgrade GeoSynchrony Procedures for each release: 6.2 and Patches
Select an upgrade procedure: Upgrade procedure

Generated: February 15, 2022 5:35 AM GMT

REPORT PROBLEMS

If you find any errors in this procedure or have comments regarding this application, send email to
[email protected]

Copyright © 2021 Dell Inc. or its subsidiaries. All Rights Reserved.

UNLESS AUTHORIZED BY EMC CORPORATION OR ONE OF ITS AFFILIATES IN WRITING THIS

DOCUMENT IS RESTRICTED FOR THE INTERNAL USE OF EMC CORPORATION (“EMC”) AND ITS
AFFILIATES INCLUDING DELL INC. THIS DOCUMENT AND THE INSTRUCTIONS CONTAINED
HEREIN ARE FOR USE BY SUCH AUTHORIZED USERS AND APPROVED EMC AND DELL SERVICE
PERSONNEL. ANY USE OF THIS DOCUMENT AND/OR ITS INSTRUCTIONS BY ANYONE OTHER
THAN AUTHORIZED USERS AND APPROVED EMC AND DELL SERVICE PERSONNEL
(“UNAUTHORIZED USER”) IS NOT AUTHORIZED, LICENSED OR PERMITTED BY EMC.
UNAUTHORIZED USER HAS EXCEEDED ANY CONTRACTUAL RIGHTS THAT MAY EXIST
BETWEEN UNAUTHORIZED USER AND EMC CORPORATION AND/OR ITS AFFILIATES, AND
ASSUMES SOLE AND COMPLETE RESPONSIBILITY AND RISK FOR ANY AND ALL DIRECT OR
INDIRECT LOSSES OR DAMAGES SUFFERED BY EMC CORPORATION, ITS AFFILIATES OR ANY
THIRD PARTIES THAT MAY ARISE OUT OF OR OCCUR IN CONNECTION WITH ANY SUCH
UNAUTHORIZED USE BY UNAUTHORIZED USER.

Dell, EMC, Dell EMC and other trademarks are trademarks of Dell Inc. or its subsidiaries. Other
trademarks may be the property of their respective owners.

Publication Date: February, 2022

Dell Technologies Confidential Information- for internal use only version: 2.9.0.87

Page 1 of 36
 Contents
Preliminary Activity Tasks .......................................................................................................3
A SolVe Engine approval is required for this activity ....................................................................................3
Read, understand, and perform these tasks.................................................................................................3

NDU_to_Geosynchrony_R6.2_Family Upgrade .....................................................................6

Dell Technologies Confidential Information- for internal use only version: 2.9.0.87

Page 2 of 36
Preliminary Activity Tasks
This section may contain tasks that you must complete before performing this procedure.

A SolVe Engine approval is required for this activity

1. Attach supporting information for the items listed in Table 1 to the Engine activity to expedite the
approval process.

Table 1 Required Information for this SolVe Engine approval

• EMPLOYEES: Please review the latest VPLEX SolVe Engine Change Control Process on the
VPLEX Wiki Change Control page to properly prepare and submit the SolVe Engine activity (Location:
https://dellservices.lightning.force.com/lightning/r/Lightning_Knowledge__kav/ka06P000000TRZIQA4/
view )
• PARTNERS: Please review the Knowledgebase Article: https://support.emc.com/kb/303768
• The SolVe Engine activity must be submitted at least 10 business days prior to the scheduled
upgrade/NDU date and must include
• The exact GeoSynchrony release version that you are upgrading from
• The exact GeoSynchrony release version that you are upgrading to
• A PuTTY log file from the cluster involved in the upgrade (maximum of 2) containing the info outlined
in the VPLEX Change Control Process document, attached to each SolVe task
• A file containing customer SAN switch details, named ‘Switch_data.txt’, which only contains the
make/model/OS version of each switch where the VPLEX is connected into the customer’s SAN
• Each cluster MUST have a separate upgrade SR and the .tar file from the output of the ‘collect-
diagnostics -–noextended’ command must be attached to the respective upgrade SR, NOT the SolVe
Engine.

2. Once approved, record the SolVe Engine activity ID here for future reference:
______________________________________________________________________________

Read, understand, and perform these tasks

3. Table 2 lists tasks, cautions, warnings, notes, and/or knowledgebase (KB) solutions that you need to
be aware of before performing this activity. Read, understand, and when necessary perform any
tasks contained in this table and any tasks contained in any associated knowledgebase solution.

Table 2 List of cautions, warnings, notes, and/or KB solutions related to this activity

• If you find an error in any VPLEX product procedure, that was generated via SolVe Desktop or SolVe
Online,you should report the issue to the SolVe Team using the feedback address located in the KBA
335501 as well surrounding the reporting of any issues,errors confusion in a VPLEX procedure

Dell Technologies Confidential Information- for internal use only version: 2.9.0.87

Page 3 of 36
 4. This is a link to the top trending service topics. These topics may or not be related to this activity.
This is merely a proactive attempt to make you aware of any KB articles that may be associated with
this product.

Note: There may not be any top trending service topics for this product at any given time.

VPLEX Top Service Topics

Dell Technologies Confidential Information- for internal use only version: 2.9.0.87

Page 4 of 36
Dell Technologies Confidential Information- for internal use only version: 2.9.0.87

Page 5 of 36
NDU_to_Geosynchrony_R6.2_Family Upgrade

Note: The next section is an existing PDF document that is inserted into this procedure. You may see
two sets of page numbers because the existing PDF has its own page numbering. Page x of y on the
bottom will be the page number of the entire procedure.

Dell Technologies Confidential Information- for internal use only version: 2.9.0.87

Page 6 of 36
VPLEX GeoSynchrony
NDU Procedure
6.2 and Patches

January 2022

Page 7 of 36
Notes, cautions, and warnings

NOTE: A NOTE indicates important information that helps you make better use of your product.

CAUTION: A CAUTION indicates either potential damage to hardware or loss of data and tells you how to avoid
the problem.

WARNING: A WARNING indicates a potential for property damage, personal injury, or death.

© 2016 - 2022 Dell Inc. or its subsidiaries. All rights reserved. Dell, EMC, and other trademarks are trademarks of Dell Inc. or its subsidiaries.
Other trademarks may be trademarks of their respective owners.

Page 8 of 36
 Contents

Chapter 1: VPLEX NDU to GeoSynchrony Releases 6.2 Family....................................................... 4

GeoSynchrony Upgrade..................................................................................................................................................... 4
NDU considerations.......................................................................................................................................................5
Using the copy and paste feature at the command prompt............................................................................... 6
Options that depend on the availability of the customer.....................................................................................6
24 Hours before beginning NDU................................................................................................................................ 7
Assess time that is required for upgrade................................................................................................................. 8
Assess readiness for write-same feature.................................................................................................................9
On the day of NDU........................................................................................................................................................9
Connect PuTTY.............................................................................................................................................................. 9
Start a new virtual terminal session........................................................................................................................ 10
Run the firmware upgrade pre-checks and health-check....................................................................................... 10
Read the Release Notes....................................................................................................................................................11
Log in with the VPLEX CLI............................................................................................................................................... 11
Verify NDU Readiness.......................................................................................................................................................12
Upgrade management servers........................................................................................................................................ 13
Disable Call Home...............................................................................................................................................................13
Upgrade the management server in cluster 1............................................................................................................. 14
Upgrade the management server in cluster 2.............................................................................................................16
Re-run NDU pre-check..................................................................................................................................................... 17
Enable Call Home................................................................................................................................................................17
Upgrade the firmware on VPLEX Local or Metro systems......................................................................................18
Verify the upgrade............................................................................................................................................................ 20
Complete the upgrade..................................................................................................................................................... 20
Steps to fix the port-stats-monitor script for post 6.2 through 6.2 Patch 3...............................................22
Exit the virtual terminal session..................................................................................................................................... 23
Appendix A: Workflow details of the I/O forwarding phase options.................................................................... 23
Appendix B: Version information................................................................................................................................... 26
NDU Worksheet.................................................................................................................................................................28

Contents 3

Page 9 of 36
 1
VPLEX NDU to GeoSynchrony Releases 6.2
Family
This document describes how to upgrade (NDU) GeoSynchrony to VPLEX 6.2 Family without disruptions.
Topics:
• GeoSynchrony Upgrade
• Run the firmware upgrade pre-checks and health-check
• Read the Release Notes
• Log in with the VPLEX CLI
• Verify NDU Readiness
• Upgrade management servers
• Disable Call Home
• Upgrade the management server in cluster 1
• Upgrade the management server in cluster 2
• Re-run NDU pre-check
• Enable Call Home
• Upgrade the firmware on VPLEX Local or Metro systems
• Verify the upgrade
• Complete the upgrade
• Exit the virtual terminal session
• Appendix A: Workflow details of the I/O forwarding phase options
• Appendix B: Version information
• NDU Worksheet

GeoSynchrony Upgrade
● Before beginning any upgrade of GeoSynchrony, read the associated Release Notes. They contain a list of fixed issues as
well as specific information such as the version number, location of files to be downloaded, upgrade paths and verification
details that are required for this upgrade. While reading the Release Notes, use the Worksheet found at the end of
this document to record the values you need for upgrading the system.
● If you are new to upgrading VPLEX, read the NDU Overview before starting an NDU. The overview describes the activities
that occur during NDU and the minimum configuration in which NDU should be run.
NOTE: NDU Overview and NDU Troubleshooting documents are available for download at SolVe.
● Failure to understand the NDU process or what constitutes a High Availability (HA) configuration can result in DU.
● This procedure requires VPLEX to be in a configured state. Specifically, the metavolume must be configured. To upgrade an
unconfigured system (a system without a metavolume), refer to the VPLEX Configuration Guide.
NOTE: This version of GeoSynchrony is a stand-alone upgrade. For example, you can upgrade from version 6.0 or later
directly to this version (including from patch releases).
The following figure provides an overview of the NDU process.

4 VPLEX NDU to GeoSynchrony Releases 6.2 Family

Page 10 of 36
Figure 1. NDU Process

NOTE: Before upgrading the GeoSynchrony version to 6.2, obtain an appropriate license file for your clusters. Install the
license after the upgrade. To obtain the proper license key file for installation, visit the Dell EMC Software Licensing Central
Website with the License Authorization Code (LAC) associated with your acquisition. VPLEX Product Guide and VPLEX
Administration Guide provide you additional details on licensing in VPLEX.

NDU considerations
To direct upgrade to version 6.2 or its patches, you must be running at least GeoSynchrony release 5.5.x or later.
To upgrade to 6.2 and patches, all VPLEX running on pre-5.5.x need one or more step upgrades. For more details, see the
Upgrading Geosynchrony section in the Release Notes 6.2 and Patches.
NOTE: Perform VS6 upgrades from MMCS-A. When you run a Management Server NDU on MMCS-A the software on
MMCS-B is upgraded as well.
For the Metro-IP systems, the following IP WAN firewall ports need to be opened before the NDU:
● TCP 61482-61484
● TCP 32768-61000
CAUTION:

Failure to open the above firewall ports will cause the NDU to fail and rollback.
Any legacy UDP ports that are opened for older VPLEX software versions can be closed after the NDU.
NOTE: In VPLEX GeoSynchrony release 5.4.1 and later, the VPLEX: call-home filename shows NOSERIALNUM instead
of the Top Level Assembly (TLA) of the local directors, which identifies the cluster where the call home event message
originates. This leads to the creation of automated SYR/CLM cases (29.AAA1.58). To resolve this issue, after an upgrade to
5.4.x and before and after an upgrade from 5.4.x to a later version, follow the procedure described in the KnowledgeBase
article 462461.
When FCIP gateways are used between the VPLEX clusters on the Metro FC VS2 and VS6 systems that run the GeoSynchrony
version 6.x, high distributed-device write latency might occur during the high distributed-device write bandwidth (MB/s)
scenario. The VS2 systems on GeoSynchrony 5.x version, which use FCIP gateways, must continue to use the 5.x version until
further notice. For more information, refer to the Knowledge Base article 527524.
If the hosts that run Red Hat Enterprise Linux (RHEL) or SUSE Linux Enterprise Server (SLES) do not have the latest version
of the Dell EMC PowerPath software (with the supported kernel version) configured on them, it might take more time for the
hosts to scan the paths and bring them alive. This issue might cause data unavailability (DU) during the GeoSynchrony upgrade.
For more information, see the Knowledge Base article 529059.

VPLEX NDU to GeoSynchrony Releases 6.2 Family 5

Page 11 of 36
Using the copy and paste feature at the command prompt
This document provides extensive examples of the required commands to perform this upgrade as well as to check system
status and configure options.
While it may be tempting to copy commands directly from the document and paste them into the command line, take care when
doing so.
In some browsers, the output from this document is interpreted incorrectly, causing an error in the command line. For example,
some single dashes (-) before flags are interpreted at the command prompt as dots (.). Some double dashes (--) are
interpreted as single dashes (-). If you must copy and paste a command from the document to the command line, check the
syntax of the command against the command in the document before entering the command.

Options that depend on the availability of the customer

The VPLEX NDU procedure now includes an option to extend the I/O forwarding phase, and a few options to determine
whether you must be prompted during the I/O forwarding phase, before proceeding with the NDU. For the successful
completion of the NDU, it is required to communicate with the customer ahead of time to determine their availability during
the NDU to assess the status of the host initiator logins during the I/O forwarding phase. Depending on the availability of the
customer during the NDU, determine which option to use for the prompt display, and how long to set the timeout setting.
The following table describes the available prompt options.

Table 1. Options for the prompt display during the I/O forwarding phase
Argument Description Prompt
--io-fwd-ask-for-confirmation After an initial wait period of 6 minutes, Continue: NDU continues even when
always the status of host initiator logins is there are missing initiator logins. Make
checked and the details are displayed. sure that the customer is aware that
You are prompted with the options at missing logins can cause DU.
this point. You can continue to type
Rollback: NDU rolls back and DU is
Refresh as needed, before the I/O
avoided. The customer can check the
forwarding phase times out.
host, resolve the issue that led to the
Choose this option if you have hosts missing initiator logins, and rerun the
that require manual scanning for the NDU.
paths to be visible. Assistance from
Refresh: Get the new list of initiators.
the customer is required to verify that
initiator paths on the hosts are alive. If If all the initiators are logged in, VPLEX
the path is unavailable, resolve the issue displays the prompts to move forward.
within the timeout period that you have
specified.

--io-fwd-ask-for-confirmation Assistance from the customer is required Continue: NDU continues even when
on-missing-logins to determine whether any missing there are missing initiator logins. Make
initiators are from critical hosts. If paths sure that the customer is aware that
are unavailable from critical hosts, the missing logins can cause DU.
customer must resolve the issue before
Rollback-: NDU rolls back and DU is
continuing with the NDU.
avoided. The customer can check the
After an initial wait period of 6 host, resolve the issue that led to the
minutes, the status of the host initiator missing initiator logins, and rerun the
logins are checked. If all the initiator NDU.
logins are detected, the NDU continues
automatically (without displaying any Refresh-: Get the new list of initiators.
prompt). If there are missing host After all the initiators are logged in, NDU
initiator logins, they are displayed, and continues without displaying any prompt.
you are prompted with the options.
You can continue to type Refresh as
needed, before the I/O forwarding phase
timeout expires.

6 VPLEX NDU to GeoSynchrony Releases 6.2 Family

Page 12 of 36
Table 1. Options for the prompt display during the I/O forwarding phase (continued)
Argument Description Prompt
--io-fwd-ask-for-confirmation No interaction is required or the This option does not create any prompt.
never customer is not available to check the After an initial wait period of 6 minutes,
host connectivity. NDU waits for all the host initiator logins are checked.
the initiators to log back in within If all initiator logins are detected, the
the specified timeout period. Resolve NDU continues automatically (without
any issues within this period. If VPLEX displaying any prompt). If there are
identifies any missing logins after the missing initiator logins, they are checked
timeout period, NDU is rolled back. again every 10 seconds automatically,
until the I/O forwarding phase times
NOTE: The default IO forwarding
out. At this point, the NDU rolls back
phase timeout is 180 minutes. It automatically (without displaying any
is important to check whether this prompt). If missing initiator logins are
value must be modified according to not detected within the timeout period,
the environment requirements of the the NDU continues.
customers and weighed against the
risks.

To understand the detailed workflow of these options, see Appendix: Workflow details of the I/O forwarding phase options.

Controlling the I/O forwarding phase timeout

During NDU, the first set of directors are booted into the target GeoSynchrony software, and when they come back up the host
initiators are expected to log back in to them. As the initiators log back in, the I/Os serviced by the first set of directors are
forwarded to the second set of directors. This is referred to as the I/O forwarding phase, and the --io-fwd-timeout option
allows the user to control the duration of this phase. By default this phase lasts for 180 minutes, but this timeout period can be
modified to a minimum of 6 minutes and a maximum of 12 hours, by using the --io-fwd-timeout= time option, where time is
the period that you want set. To specify a time, type one of the following letters after the time in numerals:
● s for seconds
● m for minutes
● h for hours
● d for days
Some customers might have hosts with special circumstances that require more time to log back in to the first set of directors
before the second set are taken down for upgrade (manual scanning might be required for this), or some customers might want
to verify that all paths are active on their critical hosts before proceeding. These circumstances warrant the use of an extended
I/O forwarding phase. However, there are significant risks involved in a prolonged I/O forwarding phase, which include:
● Performance degradation due to only half the directors processing the I/O (this is especially risky if the CPU utilization of
the directors is high before the NDU)
● Greater risk of a total cluster outage on single-engine systems (if a second upgrader director goes down during the I/O
forwarding phase)
● Greater risk of DU due to a site partition, as the Cluster Witness Server is disabled during NDU
● Greater risk of NDU rollback due to changes that require metadata updates (RecoverPoint replication failure, RAID-1 storage
accessibility changes, and so on)
● Blocking of the UI and RESTful CLI access during NDU
● Disabled state of Call-Home during the NDU
Therefore, it is recommended to use the shortest timeout that meets the needs of the customers, particularly when no
interaction is selected (--io-fwd-ask-for-confirmation = never).

24 Hours before beginning NDU

Tasks to perform twenty-four hours before starting NDU
NDU pre-checks must be run within 24 hours before starting the NDU.
Note the following before running the NDU pre-checks:

VPLEX NDU to GeoSynchrony Releases 6.2 Family 7

Page 13 of 36
● Back-end connectivity checks validate the high availability of the back-end connectivity and ensure enough redundancy to
avoid service disruption in case of failures of cabling, ports, and fabric during NDU. These checks ensure the following:
○ For active-active arrays, four paths are required for each LUN to each director. For active-passive arrays, the
requirement is four active and four passive paths for each LUN to each director for a total of eight paths.
○ There are at least two unique initiators and two unique targets in the set of paths to a LUN from each director.
● Front-end connectivity on minimal configurations does not meet high-availability requirements. Errors may be encountered
during the pre-check.
● NDU pre-checks may or may not pass for systems with partial back-end visibility, depending on the extent of the visibility
across all directors in the system. Lack of visibility may result in a firmware upgrade that fails and rolls back to the current
version.
● High latency on the management network does not meet high-availability requirements. Errors may be encountered during
the pre-check. If so, the latency on the management network should be brought beneath the 2-second threshold.
● If the NDU pre-check detects a problem in the configuration do NOT try to skip the associated pre-checks during the
firmware upgrade. Correct the system configuration error, and follow the recommended configuration according to best
practices before continuing further.

Assess time that is required for upgrade

The table outlines the approximate time that is required for the completion of each stage of the upgrade process in a VS2
system. The NDU process assesses the packages to be upgraded and determine the upgrades that are required on the director.
It also determines whether to upgrade the firmware services, the BIOS/POST, or the BIOS/POST and kernel.
The BIOS/POST upgrade times vary based on the hardware platform used. The time that is outlined is the maximum time that it
would take to complete the upgrade process.
Cluster Witness Server upgrade determines if an upgrade to the cluster witness server kernel is required. Depending on the
latency between the management-server and cluster witness server, NDU time for Cluster Witness Server upgrades vary. The
time that is outlined is maximum time.
Unless otherwise specified, all timings apply to Local and Metro with single, dual, or quad configurations.

Table 2. NDU Timings

Stages Task VS2 [Time in Minutes] VS6 [Time in Minutes]
1 Firmware upgrade pre-checks and health-check
● NDU pre-check (before upgrade) 10 10
● System health-check (contingent on 10 10
number of volumes)
2 Management server upgrade
● Local 20 20
● Metro (for single, dual, and quad 40 40
configurations)
3 NDU pre-checks (after upgrade) 10 10
4 GeoSynchrony upgrade (NDU) [includes 40 a 90 b

BIOS/POST and kernel upgrade]

5 NDU pre-check (after NDU) 10 10
6 VPLEX Witness Upgrade 60 60
Total Upgrade Time in Minutes 200 250

a. This does not include the 180 minutes, which is the default timeout period of the I/O forwarding phase. You can set this
timeout period to a minimum of 6 minutes and a maximum of 12 hours. Controlling the I/O forwarding phase timeout
provides you additional information on the I/O forwarding phase timeout.
b. This does not include the 180 minutes, which is the default timeout period of the I/O forwarding phase. You can set this
timeout period to a minimum of 6 minutes and a maximum of 12 hours. Controlling the I/O forwarding phase timeout
provides you additional information on the I/O forwarding phase timeout.

8 VPLEX NDU to GeoSynchrony Releases 6.2 Family

Page 14 of 36
Assess readiness for write-same feature
Before upgrade, ensure that the arrays are configured to handle the write-same feature.
For AMS 25xx and HUS 110/150 arrays, refer to the Configuring Arrays procedure in the SolVe Desktop generator for steps to
set these arrays to support the write-same feature. For all other arrays, follow the documentation that is provided by the array
vendor to enable the write-same feature for VPLEX.

On the day of NDU

Tasks to perform on the day of running NDU
Before starting the NDU, record the information from the Collect Information section in the NDU worksheet at the end of this
document. You need this information throughout the procedure.
After filling out the NDU Worksheet, continue to the next section of this document.
Additionally, configure all the RecoverPoint consistency groups to use asynchronous protection. To do this, navigate to the Link
Policy tab under Protection > Manage Protection in the RecoverPoint GUI.

Connect PuTTY
Connect to the management server using PuTTY
Use either of the following methods to connect to the management server:
● A service laptop that is connected to the service port on the management server.
● PuTTY or SSH on any computer able to connect to the public IP address of the management server.
NOTE: Verify that SSH or PuTTY is set to version 2.

1. Start PuTTY.
2. Connect with the username service.
For VPLEX Metro systems, start two connections (one to each management server).
In this example, PuTTy is configured to use SSH version 2:

PuTTY is connected via the management service.

VPLEX NDU to GeoSynchrony Releases 6.2 Family 9

Page 15 of 36
Start a new virtual terminal session
To preserve the session from an SSH session loss, start a new virtual terminal session:
screen -S session_name
Where session_name is the name of the session, as displayed in the following example:

service@ManagementServer:~> screen -S upgrade

NOTE: This virtual terminal session is deleted after the Management Server restarts as part of NDU process. To continue
with NDU after the Management Server restarts, start a new session. To manage your sessions, use these commands:
● screen ls - To view all the running sessions.
● screen -dR session_name - To recover a session at any point .

It is recommended to use a virtual terminal session when a longer timeout period is set for the I/O forwarding phase. Controlling
the I/O forwarding phase timeout provides you additional information on setting the I/O forwarding phase timeout.

Run the firmware upgrade pre-checks and health-

check
Run firmware upgrade pre-checks and health-check to ensure a successful NDU. You can perform the NDU pre-checks from
the management server in either cluster. However, this procedure assumes that you are logged into the management server
on cluster 1. The default password is located in the VPLEX Security Configuration Guide. If the customer has changed the
password, contact the customer for the new password.
1. On cluster 1, log in to the VPlexcli with the username service.

service@ManagementServer:~> vplexcli
Trying ::1...
Connected to localhost.
Escape character is '^]'.

Enter User Name: service

Password: <service password>

creating logfile:/var/log/VPlex/cli/session.log_service_localhost_Txxxx_YYYYMMDDHHMMSS

Where service password is the password of your service account.

NOTE: This example is from a version before 6.0.

2. VPlexcli At the VPlexcli prompt, type the health-check --full command to verify that all values display OK.

VPlexcli:/> health-check --full

3. Verify that all checks return a status of OK. If errors are displayed, resolve the errors before continuing with the NDU.
Consult the NDU Troubleshooting section or contact Dell EMC Customer Support. When all errors are corrected, use the
health-check --full command again to verify that all errors are resolved.
NOTE: Do NOT continue until all checks return a status of OK.

4. VPlexcli At the VPlexcli prompt, type the ndu pre-check command to ensure that the system is healthy and ready for
NDU.

VPlexcli:/> ndu pre-check

Warning:
During the NDU process, multiple directors will be offline for a
portion of the time. This is non-disruptive but is dependent on
a host-based multipathing solution being installed, configured,
and operating on all connected hosts.

Warning:

10 VPLEX NDU to GeoSynchrony Releases 6.2 Family

Page 16 of 36
 Please run the "health-check --full" command to verify that the VPLEX
hardware is healthy.
Analyzing system configuration: .DONE
===========================================================================
Performing NDU pre-checks
===========================================================================
Verify NDU is not in progress.. OK
Verify director communication status.. OK
Verify management network redundancy.. OK
.
.
.
Verify cluster witness state.. OK

==========================================================================

==========================================================================
No problems found during NDU pre-check.
==========================================================================
The output for 'ndu pre-check' has been captured
in /var/log/VPlex/cli/capture/ndu-pre-check-session.txt

VPlexcli:/>

5. Verify that all checks return a status of OK.

6. Verify that the message No problems found during NDU pre-check appears in the output. If the message does
not appear, or if errors are displayed, resolve the errors before continuing with the NDU. Consult the NDU Troubleshooting
section or contact Dell EMC Customer Support. When all errors are corrected, use the NDU pre-check command again to
verify that all errors are resolved.
NOTE: Do NOT continue until the No problems found during NDU pre-check message is displayed. To resolve any
issues, refer to the troubleshooting information available at Installation and Upgrade> Upgrade GeoSynchrony >
Upgrade GeoSynchrony 6.2 > NDU Troubleshooting in SolVe.
7. Verify that all hosts are configured with multipath or failover drivers, and verify that the multi-pathing configuration is in
good standing with no path issues.

Read the Release Notes

Read the Release Notes for the release you are upgrading to. Collect the following information from the release notes and
record it in the NDU worksheet in the Version Number from the Software Version section:
● The Name and location of the management server tar package
● The name and location of the director firmware package
● The name and location of the management server ISO package
NOTE: Read the "Known Issues" section in the Release Notes and ensure that you take the necessary actions before
performing the NDU.

Log in with the VPLEX CLI

Log in to the VPLEX CLI prompt by doing the following:
1. Log in to the VPLEX management server by typing its IP address in PuTTY. The default password is listed in the VPLEX
Security Configuration Guide. However, this password might have been changed. Contact the customer to obtain the
current service password. login as:
a. Login as service using keyboard-interactive authentication.
b. Type the password.
2. Log in to VPlexCLI by typing vplexcli and providing the VPlexCLI username and password.
Sample output follows:

service@vplexname:~> vplexcli
Trying 127.0.0.1...
Connected to localhost.
Escape character is '^]'.

VPLEX NDU to GeoSynchrony Releases 6.2 Family 11

Page 17 of 36
 Enter User Name: service
Password:
creating logfile:/var/log/VPlex/cli/session.
log_service_localhost_T01234_20110525063015
VPlexcli:/>

Verify NDU Readiness

Perform the following tasks to verify that VPLEX, the host environment, and connectivity between clusters (for Metro
configurations) are ready for upgrade.
1. Verify the product version.
a. In the session to the management server on cluster 1, type the version command to display the product version.

VPlexcli:/ version
What Version Info
-------------------- --------------- ----
Product Version version -
SMSv2 D30.60.0.3.0 -
Mgmt Server Base D30.0.0.112 -
Mgmt Server Software D30.60.0.3 -

b. The value that is displayed in the GeoSynchrony version field reflects the version that you are upgrading from. Record
this product version number in the NDU Worksheet. You need this information later in the procedure.
c. See the Dell EMC VPLEX Release Notes to verify that the product version is one of the versions that are listed in the
Upgrade Paths section.
d. For VPLEX Metro systems, log in to the VPlexcli on the management server on cluster 2 and verify that the
GeoSynchrony versions of cluster 1 and cluster 2 match.
2. Verify Cluster Witness (for Metro only).
a. VPlexcli At the VPlexcli prompt (on either cluster), type the following command to check if Cluster Witness is
configured:

ll /cluster-witness/

If VPLEX Witness is not configured, the output of the ll command displays No context found for '/cluster-
witness/.

VPlexcli:/ ll/cluster-witness/
ls: No context found for '/cluster-witness/'

b. Record the VPLEX Witness configuration status in the NDU Worksheet.

c. If it is configured, make a note of its public IP address in the NDU Worksheet.
3. Verify if RecoverPoint splitters are present.
If you are running RecoverPoint, the splitter is upgraded as part of VPLEX NDU.
During NDU, RecoverPoint Appliances (RPAs) are put into maintenance mode. Maintenance Mode prevents any
RecoverPoint configuration changes from being made during NDU.
RPAs can be upgraded either before or after the VPLEX splitter is upgraded (by VPLEX NDU), and can support VPLEX
splitter versions with at least one version of discrepancy (N-1, N and N+1).
● See the Dell EMC RecoverPoint Administration Guide for information about upgrading RPAs.
● See the Dell EMC Simple Support Matrix for information about supported versions.
● See the Dell EMC VPLEX Release Notes for both VPLEX and RecoverPoint and ensure that you take appropriate steps
regarding RecoverPoint issues.

a. VPlexcli At the VPlexcli prompt, type the following to check if any RecoverPoint clusters are configured:

ll /recoverpoint/rpa-clusters

12 VPLEX NDU to GeoSynchrony Releases 6.2 Family

Page 18 of 36
 If a RecoverPoint cluster is not configured, the output shows no RecoverPoint clusters:

VPlexcli:/> ll recoverpoint/rpa-clusters/
/recoverpoint/rpa-clusters:

b. Record the RecoverPoint configuration status in the NDU Worksheet.

c. If it is configured, make note of the RPA Host IP address.
d. Run the rp validate-configuration command to verify that RecoverPoint is healthy. Output similar to the
following is displayed: Your system appears healthy. There were no problems found in the
system.
e. Verify that the system is healthy. If any issues are found, fix them before continuing.
4. Verify the host environment. Verify that all host multipath applications are enabled and operating correctly.
Applications should be set as follows.
● HP-UX: PVLinks set to failover. Consult Dell EMC Host Connectivity Guide for HP-UX.
● VMware: Native MPIO set to Round Robin. Consult the VPLEX Host Multipathing: Implementation Planning and Best
Practices white paper for details.
● IBM AIX: Native MPIO set to Round Robin. Consult Dell EMC Host Connectivity Guide for IBM AIX.
● Linux: MPIO set to Round Robin Load Balancing. Consult Dell EMC Host Connectivity Guide for Linux.
● All platforms: PowerPath set to Adaptive.
NOTE: Additional information is provided in the VPLEX Host Multipathing: Implementation Planning and Best Practices
white paper on Dell EMC Support Online at: https://support.emc.com.

5. Verify system health.

At the VPlexcli prompt, type the health-check --full --verbose command to verify that all status values are OK:

VPlexcli:/> health-check --full --verbose

For minimal configurations, the No Dual Paths column indicates that high-availability requirements are not met. If a
metavolume and metavolume backup are not present, the system is not fully configured. Consult the VPLEX Configuration
Guide for the steps to complete the install and configuration. Do not continue if a metavolume and metavolume backup are
not present.
NOTE: Resolve all issues before continuing to the next step. Consult the NDU Troubleshooting procedure in the SolVe
Desktop or contact Dell EMC Customer Support.
The next steps vary depending on the number of clusters in the system:
● For single-cluster systems, skip to Upgrade management servers.
● For VPLEX Metro, continue to the next section and verify connectivity between clusters.

Once complete, the system is ready for NDU.

Upgrade management servers

Upgrade the management servers using the following procedures.
NOTE: Perform VS6 upgrades from MMCS-A. When you run a Management Server NDU on MMCS-A the software on
MMCS-B is upgraded as well.

Disable Call Home

Disable call-home, to prevent call-homes during the remainder of this procedure.
1. VPlexcli At the VPlexcli prompt, type the following command to browse to the call-home context:

VPlexcli:/> cd notifications/call-home/

VPLEX NDU to GeoSynchrony Releases 6.2 Family 13

Page 19 of 36
2. Type the following command to check the value of enabled attribute.

VPlexcli:/notifications/call-home> ls
Attributes:
Name Value
------- -----
enabled true
Contexts:
snmp-traps

If the enabled property value is false, do not perform the next step.
Note whether call home was enabled or disabled. Later in the procedure, you need this information to determine whether to
enable call home again.

3. Type the following command to disable call home:

VPlexcli:/notifications/call-home> set enabled false --force

If this command worked, an ls of the context shows that the enabled state of the call home is false .

Upgrade the management server in cluster 1

If you are upgrading a VPLEX Local system, upgrade the management server in cluster 1. Do not upgrade the management
server in cluster 2 or verify remote directors. After completing this task, continue directly to the Enable Call Home task.
If you are upgrading a VPLEX Metro system, complete the steps in this task, and then complete the steps as described in
Upgrade the management server in cluster 2.

1. VPlexcli At the VPlexcli prompt on cluster 1, type the exit command to exit VPlexcli:

VPlexcli:/> exit
Connection closed by foreign host.

service@ManagementServer:~>

2. At the management server prompt, type the following command to change the current directory to /tmp/
VPlexInstallPackages, as in this example:

service@ManagementServer:~> cd /tmp/VPlexInstallPackages
service@ManagementServer:/tmp/VPlexInstallPackages>

NOTE: The management server upgrade must be run from the /tmp/VPlexInstallPackages directory.

3. Copy the management server package for the new GeoSynchrony release from the site that is documented in the Release
Notes (and recorded in the NDU Worksheet) to the /tmp/VPlexInstallPackages directory on the management server.
NOTE: Do not upload the director firmware tar files at this step. The management server install process removes all files
from the /tmp/VPlexInstallPackages directory.

For example:

service@ManagementSever:/tmp/VPlexInstallPackages> scp -p <username>@<ip_address>:/

VPLEX_builds/VPLEX-version/VPLEX_PACKAGES/VPlex-version-management-server-package.tar
/tmp/VPlexInstallPackages

Where:
● username is the username on the server where the management server package has to be copied from.
● ip_address is the IP address of the server where the management server package has to be copied from.
4. At the management server prompt, type ls tar-file-name to confirm that the packages are present:

14 VPLEX NDU to GeoSynchrony Releases 6.2 Family

Page 20 of 36
 service@ManagementServer:/tmp/VPlexInstallPackages> ls VPlex-version-management-server-
package.tar
VPlex-version-management-server-package.tar
5. Retrieve the management server software package SHA-256 checksum value from the site that is documented in Release
Notes (recorded in the NDU Worksheet).

service@ManagementServer:/tmp> sha256sum VPlex-version-management-server-package.tar

6. Type the following command to run VPlex-MS-installer to install the new management server package:

VPlex-MS-installer -v VPlex-version-management-server-package.tar

NOTE: The management server upgrade may restart the management server.

A series of messages is displayed. For example:

service@sms-dolce:/tmp/VPlexInstallPackages> VPlex-MS-installer -v VPlex-version-

management-server-package.tar
Checking whether the upgrade path is supported

NOTE: Contact Dell EMC Customer Support if this command fails with an error.

7. Type the password for the service account, if prompted. A series of messages is displayed (The following example is from
the version 6.2. The management service package to be used must reflect the version to which the system is upgraded to,
for example, VPlex-6.2.x.xx.xx.xx- management-server-package.tar).

service@dualenginelocal:/tmp/VPlexInstallPackages> VPlex-MS-installer -v
VPlex-6.2.0.00.00.32-management-server-package.tar
Verify that the management server version is the same on both MMCSs.
ok
Checking whether the upgrade path is supported
Enter the service account password:

ok
Upgrade peer MMCS before local MMCS. Transfer /tmp/VPlexInstallPackages to peer MMCS.
ok

Cleaning old filesystem snapshots

ok
Upgrading software packages
ok

NOTE: If the service@sms-dolce message appears, the system was rebooted

8. VPlexcli In the VPlexcli, type the version command to confirm that the versions of the installed management server
package match the versions in the output:
For GeoSynchrony 6.2:

VPlexcli:/> version
What Version Info
------------------------------- -------------- ----------------------------------
Product Version 6.2.0.00.00.32 -
SMSv2 162.0.0.324.0 -
Mgmt Server Base 162.0.0.240 -
Mgmt Server Software 162.0.0.422 -
Cluster Witness Server Software 162.0.0.422 Built against
GeoSynchrony version - 162.0.0.422

During management server upgrade and NDU, output of the version command displays the message: Version
mismatch (or NDU) Version on one or more directors does not match mgmt server or an NDU
is in progress. This message is expected and can be ignored.

VPLEX NDU to GeoSynchrony Releases 6.2 Family 15

Page 21 of 36
9. If you are upgrading a VPLEX Local system, note that local upgrades only require an upgrade of the management server in
cluster 1. After completing this task, continue directly to Enable Call Home.
Upgrade of the management server on cluster 1 is complete.

Upgrade the management server in cluster 2

1. VPlexcli At the VPlexcli prompt on cluster 2, type the exit command to exit VPlexcli:

VPlexcli:/> exit
Connection closed by foreign host.

service@ManagementServer:~>

2. At the management server prompt, type the following command to change the current directory to /tmp/
VPlexInstallPackages, as in this example:

service@ManagementServer:~> cd /tmp/VPlexInstallPackages
service@ManagementServer:/tmp/VPlexInstallPackages>

NOTE: The management server upgrade must be run from the /tmp/VPlexInstallPackages directory.

3. Copy the management server package for the new GeoSynchrony release from the site that is documented in the Release
Notes (and recorded in the NDU Worksheet) to the /tmp/VPlexInstallPackages directory on the management server.
NOTE: Do not upload the director firmware tar files at this step. The management server install process removes all files
from the /tmp/VPlexInstallPackages directory.

For example:

service@ManagementSever:/tmp/VPlexInstallPackages> scp -p <username>@<ip_address>:/

VPLEX_builds/VPLEX-version/VPLEX_PACKAGES/VPlex-version-management-server-package.tar
/tmp/VPlexInstallPackages

Where:
● username is the username on the server where management server package has to be copied from.
○ ip_address is the IP address of the server where management server package has to be copied from.
4. At the management server prompt, type ls tar-file-name to confirm that the packages are present:
service@ManagementServer:/tmp/VPlexInstallPackages> ls VPlex-version-management-server-
package.tar
VPlex-version-management-server-package.tar
5. Retrieve the management server software package SHA-256 checksum value from the site that is documented in Release
Notes (recorded in the NDU Worksheet).

service@ManagementServer:/tmp> sha256sum VPlex-version-management-server-package.tar

6. Type the following command to run VPlex-MS-installer to install the new management server package:

VPlex-MS-installer -v VPlex-version-management-server-package.tar

NOTE: The management server upgrade may restart the management server.

A series of messages is displayed. For example:

service@sms-dolce:/tmp/VPlexInstallPackages> VPlex-MS-installer -v VPlex-version-

management-server-package.tar
Checking whether the upgrade path is supported

16 VPLEX NDU to GeoSynchrony Releases 6.2 Family

Page 22 of 36
 NOTE: Contact Dell EMC Customer Support if this command fails with an error.

7. Type the password for the service account, if prompted. A series of messages is displayed (The following example is from
the version 6.2. The management service package to be used must reflect the version to which the system is upgraded to,
for example, VPlex-6.2.x.xx.xx.xx- management-server-package.tar).

service@dualenginelocal:/tmp/VPlexInstallPackages> VPlex-MS-installer -v
VPlex-6.2.0.00.00.32-management-server-package.tar
Verify that the management server version is the same on both MMCSs.
ok
Checking whether the upgrade path is supported
Enter the service account password:

ok
Upgrade peer MMCS before local MMCS. Transfer /tmp/VPlexInstallPackages to peer MMCS.
ok

Cleaning old filesystem snapshots

ok
Upgrading software packages
ok

NOTE: If the service@sms-dolce message appears, the system was rebooted

8. VPlexcli In the VPlexcli, type the version command to confirm that the versions of the installed management server
package match the versions in the output:
For GeoSynchrony 6.2:

VPlexcli:/> version
What Version Info
------------------------------- -------------- ----------------------------------
Product Version 6.2.0.00.00.32 -
SMSv2 162.0.0.324.0 -
Mgmt Server Base 162.0.0.240 -
Mgmt Server Software 162.0.0.422 -
Cluster Witness Server Software 162.0.0.422 Built against
GeoSynchrony version - 162.0.0.422

During management server upgrade and NDU, output of the version command displays the message: Version
mismatch (or NDU) Version on one or more directors does not match mgmt server or an NDU
is in progress. This message is expected and can be ignored.

9. If you are upgrading a VPLEX Local system, note that local upgrades only require an upgrade of the management server in
cluster 1. After completing this task, continue directly to Enable Call Home.
Upgrade of the management server on cluster 2 is complete.

Re-run NDU pre-check

After upgrading the management server, re-run the ndu pre-check command to verify any new pre-check items that may
have been added as part of the management server upgrade.

Enable Call Home

Do the following to enable Call Home.
1. VPlexcli At the VPlexcli prompt, type the following to enable Call Home:

VPlexcli:/>cd /notifications/call-home
VPlexcli:/notifications/call-home> set enabled true

VPLEX NDU to GeoSynchrony Releases 6.2 Family 17

Page 23 of 36
2. Verify Call Home has been enabled by typing the ls command.
If Call Home is enabled, the following output should appear.

Attributes:
Name Value
---- -----
enabled true

Call Home is enabled.

Upgrade the firmware on VPLEX Local or Metro

systems
Note the following before starting the NDU:
● This section describes upgrade on VPLEX Local and Metro systems only.
● The GUI is inaccessible during NDU. Do NOT run any VPlexcli commands during NDU.
● While the firmware upgrade is in progress, call home is automatically disabled and re-enabled.
● On VPLEX Metro systems, the NDU command upgrades BOTH clusters.
● The NDU may take several minutes to complete, and includes rebooting the directors. To estimate the timings that are
required for the various stages in the NDU, refer to the section "Assess time that is required for the upgrade."
NOTE: There is no support for VPLEX Geo systems.

Perform the following to upgrade firmware on VPLEX Local and Metro systems.
1. Login to the management server on cluster-1.
2. Copy the director firmware package from the site that is documented in the Release Notes (and recorded in the NDU
Worksheet) to the /tmp/VPlexInstallPackages directory on the management server.
3. At the management server prompt, type ls *director* to confirm that the package is present.

service@ManagementServer:->cd /tmp/VPlexInstallPackages/
service@ManagementServer:/tmp/VPlexInstallPackages> ls *director*
VPlex-version-director-firmware-package.tar

4. Retrieve the director firmware software package SHA-256 checksum value from the site that is documented in the Release
Notes (that should be recorded in the NDU Worksheet).
5. Verify that the SHA-256 checksum of the director firmware package on the management server matches that retrieved in
the NDU Worksheet by typing the following.

service@ManagementServer:/tmp/VPlexInstallPackages> sha256sum VPlex-version-director-

firmware-package.tar

If the checksum does not match, remove the corrupted director firmware package and return to step 4.
6. VPlexcli Login to the VPlexcli with the username service and associated password.

7. VPlexcli At the VPlexcli prompt (on cluster 1), type the following command:

ndu start --io-fwd-ask-for-confirmation prompt type --io-fwd-timeout time -u path

Where:
- prompt type is the type of the prompt that you want to see during the IO forwarding phase of the NDU. Type:

● always - Choose this option if you have hosts that require manual scanning for the paths to be visible. Assistance from
the customer is required to verify that initiator paths on the hosts are alive. If the path is unavailable, resolve the issue
within the timeout period that you have specified. The prompts for this options are:
○ Continue: NDU continues even when there are missing initiator logins. Make sure that the customer is aware that
missing logins can cause DU.

18 VPLEX NDU to GeoSynchrony Releases 6.2 Family

Page 24 of 36
 ○ Rollback: NDU rolls back and DU is avoided. The customer can check the host, resolve the issue that led to the
missing initiator logins, and rerun the NDU.
○ Refresh: Get the new list of initiators. If all the initiators are logged in, VPLEX displays the prompts to move
forward.
● on-missing-logins - Assistance from the customer is required to determine whether any missing initiators are from
critical hosts. If paths are unavailable from critical hosts, the customer will need to resolve the issue before continuing
with the NDU. The prompts for this options are:
○ Continue: NDU continues even when there are missing initiator logins. Make sure that the customer is aware that
missing logins can cause DU.
○ Rollback: NDU rolls back and DU is avoided. The customer can check the host, resolve the issue that led to the
missing initiator logins, and rerun the NDU.
○ Refresh: Get the new list of initiators. After all the initiators are logged in, NDU continues without displaying any
prompt.
● never - No interaction is required or the customer is not available to check the host connectivity. NDU waits for all the
initiators to log back in within the specified timeout period. Resolve any issues within this period. If VPLEX identifies any
missing logins after the timeout period, NDU is rolled back. It is important to check whether this value must be modified
according to the environment requirements of the customers and weighed against the risks.
For more information on these options, see Options for the prompt display during the IO forwarding phase.
time is the period after which the I/O forward phase times out. In the I/O forward phase, the I/Os that are serviced to the
first set of directors are forwarded to the second set of directors. The hosts are expected to connect back to the first set of
directors during this period. By default, this phase lasts for 180 minutes. You can set this timeout period to a minimum of 6
minutes and a maximum of 12 hours. To specify a time, type one of the following letters after the time in numerals:
● s for seconds
● m for minutes
● h for hours
● d for days
Controlling the I/O forwarding phase timeout provides you additional information on the options.
- path is the complete path to the director software package.
.
The following is an example:

ndu start --io-fwd-ask-for-confirmation always --io-fwd-timeout 1h -u /tmp/

VPlexInstallPackages/VPlex-version-director-firmware-package.tar
===============================================================================
WARNING: IO Forwarding period timeout has been set to 60 minutes. Running the VPLEX
in the IO forwarding state for a long period
of time is subject to many risks, including the following:
* Performance degradation due to only half the directors processing IO (this is
especially risky if the directors' CPU utilization
is already high prior to NDU).
* Greater risk of a Total Cluster Outage on single-engine systems (if a second
upgrader director goes down during
the IO forwarding phase).
* Greater risk of DU due to a site partition, as the Cluster Witness Server is
disabled during NDU.

To accept the risks and continue with the specified io-fwd-timeout enter 'CONTINUE'.
Otherwise enter 'EXIT' to exit now
and start over: CONTINUE
Initializing VPLEX NDU state: ......DONE

================================================================================
[Tue Jul 10 08:37:34 2018] Preparing for NDU
================================================================================

8. When the upgrade completes, the NDU Finished Successfully message appears. If you do not see this message,
consult the Troubleshooting section or contact Dell EMC Customer Support.

VPLEX NDU to GeoSynchrony Releases 6.2 Family 19

Page 25 of 36
Verify the upgrade
Perform the following tasks to verify the upgrade was successful.
1. VPlexcli At the VPlexcli prompt, type the version -a --verbose command to display version information.
To view the output of this command in various versions, see Appendix B: Version information.
2. Verify that the product version is the version you are upgrading to, and the product component versions are as shown in the
example version output.
NOTE: If you are installing a patch release, see "Version Numbers of Components in a Patch Release" for the specific
version numbers you should see displayed in this output.

3. Verify that the OS version on all directors matches the OS version that is listed in the Release Notes.
4. Verify that the SSD Firmware Version Number on all directors meets the specification that is defined in the following
table. If the Version Number is lower than the specification, reference "Upgrade SSD Firmware" in the NDU Overview and
Background document, available on the SolVe Desktop.

Table 3. SSD Firmware Version Numbers for Directors Per Platform

Platform Vendor Model Number Dell EMC Part Number SSD Firmware Version
Number (Minimum required)
VS2 M8SB1-56UC-EMC 118032769 C08-7141
VS2 M8SB1-56UC-EMC-H 118032821 C08-7139
VS2 P30056-MTFDBAA056SAL 118032803 0005
VS6 SHM2S6D064GHM11EMC 118000106 S8FM08.0

5. Verify the state of Call Home. It is especially important to ensure that Call Home is enabled at the end of an upgrade.

cd /notifications/call-home

ls

Attributes:
Name Value
---- -----
enabled true

If the value shown is Enabled false, type the set enabled true command.

A successful upgrade has been verified.

Complete the upgrade

Perform the following tasks to complete the NDU.
1. VPlexcli At the CLI prompt, type the health-check --full command to verify cluster status, metavolume status,
front-end and back-end status. Resolve any issues from the health-check before continuing.
For minimal configurations, check the No Dual Paths column. If there is a value greater than zero in this column, this
cluster has not met the requirement.
If health-check detects that Call Home is disabled, perform the NDU task "Enable Call Home'" in the "Upgrade
management servers" section of the NDU procedure.
2. Check the rebuild status (VPLEX Metro only).
VPlexcli
At the VPlexcli prompt, type the rebuild status command to display whether any log rebuilds are in progress.

20 VPLEX NDU to GeoSynchrony Releases 6.2 Family

Page 26 of 36
 The following example shows a rebuild in progress:

VPlexcli:/> rebuild status

[10] storage_volumes marked for rebuild

Global rebuilds:
device rebuild type rebuilder director rebuilt/total percent throughput ETA
--------------- ------------ ------------------ ------------- ------- ----------
---------
cx4_dr1_0 logging s1_3b201eb7_spa 384M/2G 18.75% 16.7M/s
8s
cx4_dr1_1 logging s1_3b301eb7_spb 384M/2G 18.75% 11.2M/s
13s
cx4_dr1_3 logging s1_3b301eb7_spb 288M/2G 14.06% 10.8M/s
0.3min
cx4_dr1_4 logging s1_3b201eb7_spa 160M/2G 7.81% 17.8M/s
23s
dr1_x_wins logging s1_3b201eb7_spa 160M/10G 1.56% 14.8M/s
3.15min

Local rebuilds:
No active local rebuilds.

Wait until all rebuilds complete and output from the command appears as follows:

VPlexcli:/> rebuild status

Global rebuilds:
No active global rebuilds.

Local rebuilds:
No active local rebuilds

3. Collect SYR configuration.

VPlexcli At the VPlexcli prompt, type the syrcollect command to send an SYR configuration file to Dell EMC.
NOTE: Run the syrcollect command with no parameters. It can take several minutes to collect the information and
send the file.

4. Return RecoverPoint consistency groups to Synchronous Mode.

If the RecoverPoint consistency groups were configured to use asynchronous protection before starting the NDU, change
the configuration to use synchronous protection. To do this, navigate to the Link Policy tab under Protection > Manage
Protection in the RecoverPoint GUI.
5. Verify password status.
After upgrade, reset the passwords of existing custom local users with the user reset -u command as documented in
the VPLEX Security Configuration Guide. Do not reset default user accounts such as service and admin. It also does not
apply to LDAP/AD managed user accounts.
CAUTION: The service account password must be updated following upgrade. The service account passwords
only expire after a ten-year period, so it is essential they be changed for security purposes. The following
password policies are enforced for administrator and user accounts that are created on the management
server. These policies do NOT apply to the service user account.
● Passwords must be minimum eight characters in length.
● Passwords must be changed every 90 days.
● Passwords cannot be changed within one day of the last change.
● 75 days after the last password change, GeoSynchrony sends a reminder that the password is about to expire. This
reminder is displayed after the user has successfully logged in.
● There is a one day grace period 90 days after the last password change.
● The password expires and the account is locked, 91 days after the last password change.
NOTE: This password policy does not apply to the service user account. Therefore, the service password does not
expire after 91 days, but rather, after ten years. It is recommended strongly to change the service user account
password immediately following upgrade.

VPLEX NDU to GeoSynchrony Releases 6.2 Family 21

Page 27 of 36
 NOTE: This version of GeoSynchrony supports Role-based Access. Refer to the section "Role-based Access Control
and NDU" in the Dell EMC VPLEX Security Configuration Guide for information about how security policies affect shell
access.

6. Verify password status on VASA Provider or RecoverPoint appliance, as needed.

After upgrade, update the password on the VASA Provider and/or RecoverPoint appliance if you are using the
administrator account or another user account to connect to the management server and the password of this account is
changed on the management server.
Refer to the following documents for information on changing the passwords:
● VASA Provider appliance: Refer to "Re-configure connection to VPLEX from VASA Provider" in the NDU Overview and
Background document, as well as other VASA Installation procedures available from SolVe Desktop.
● RecoverPoint appliance: Refer to the instructions in the RecoverPoint Deployment Manager Product Guide
for updating the VPLEX splitter credentials and importing the VPLEX splitter certificate.
NOTE: If the password for the administrator account is not available or if the password has expired, contact Dell EMC
Customer Support to reset the password.

7. Set the OS2007 bit on Symmetrix arrays, as needed.

Since release 5.2, VPLEX supports the OS2007 bit on Symmetrix arrays. This setting is vital to detect LUN swap situations
and storage volume expansion automatically on a Symmetrix array. The Symmetrix section of the Configuring Arrays
procedure in the SolVe Desktop contains instructions on setting the OS2007 bit.
8. Enable the fast-write feature.
If you disabled the FAST® write feature on the FCIP WAN switches connected to VPLEX before starting the NDU, enable
the FAST® write feature now that NDU has completed.
To enable the FAST® write feature on a switch, refer to the Cisco or Brocade FCIP switch administrator's guide.

9. Login to the web GUI and verify that the login credentials are working and that the GUI is populated appropriately.
10. (For the VS2 systems) If you are upgrading from a release before Release 5.5 SP1, enable the thin-aware feature on storage
volumes that support Thin Provisioning.
For information on which arrays support Thin Provisioning on VPLEX, see the Simple Support Matrix for VPLEX on Dell EMC
Support Online.
Use the virtual-volume set-thin-enabled true --virtual-volumes virtual-volumes command to enable
thin aware features. VPLEX recognizes the volumes as thin-capable during discovery. By default, they are not set to
thin-enabled. See the VPLEX Administration Guide for details of provisioning and configuring Thin Aware storage volumes.

11. Install an appropriate license for your configuration. Refer to the VPLEX Administration Guide for more information on
installing the license file.

Steps to fix the port-stats-monitor script for post 6.2 through 6.2
Patch 3
This section covers the steps to replace the corrupted script that is installed during the NDU.
NOTE: Do not use this section for 6.2 P4 and later. After upgrading to 6.2 P4, see the section STEPS TO REMOVE THE
TWO LINES ADDED IN THE WORKAROUND ONE ABOVE in the KBA 000157798 to run the script.
VPLEX that is running GeoSynchrony 6.2.x, the port monitoring script is already installed and enabled by default, yet it has been
discovered that there is a memory leak with the current python implementation portion of the script.
Workaround for 6.2.x port monitor script leak issue:
1. Download a copy of the script port_stats_monitor located on the Dell Support site under VPLEX Series > Drivers
and Downloads.
2. Uncompress the .zip file. There are three files, the script and two .json files: port-stats-moitor.py, vs2_config.json,
vs6_config.json.
3. Upload the new version of the script port_stats_monitor.py to the management server folder /opt/emc/VPlex/
jython<tab>/LibExt

22 VPLEX NDU to GeoSynchrony Releases 6.2 Family

Page 28 of 36
 NOTE: /opt/emc/VPlex/jython<tab>/, <tab> it means that press the Tab key after typing jython to auto enter the
correct jython folder name.
Sample of directory string after tabbing for jython folder on the VPLEX the script is being loaded onto:

service@ManagementServer:~> cd /opt/emc/VPlex/jython2.7b4/LibExt
service@ManagementServer:/opt/emc/VPlex/jython2.7b4/LibExt>

4. Restart the VPLEX Management console. Example:

service@ManagementServer:~> sudo /etc/init.d/VPlexManagementConsole restart

Restarting EMC VPlex Management Console... done

5. Repeat the steps 1 to 4 on the peer management server if it is a Metro.

6. After restarting the VPlex Management Console, wait for 5 minutes, and then access the VPlexcli.
7. From VPlexcli prompt, to check the status of the port monitoring script, run the command port-monitor status.

Example:
VPlexcli:/> port-monitor status
Status: running with the following parameters:
Emails: None
SMTP: None
Local-only: False
Threshold config: None
Directors: director-1-1-B, director-2-1-A, director-1-1-A,
director-2-1-B, director-1-2-A, director-2-2-A, director-1-2-B,
director-2-2-B

NOTE: To configure the port monitoring script to send email, or if thresholds are required to be adjusted, the field
has to work with the customer to do it. The steps for doing it are available in the Dell EMC VPLEX Geosynchrony
Administration Guide 6.2 under Performance and Monitoring.

Exit the virtual terminal session

Type the following command to exit the virtual terminal session:
exit

service@ManagementServer:~> exit

Appendix A: Workflow details of the I/O forwarding

phase options
The following illustrates explains the workflow when the --io-fwd-ask-for-confirmation always option is selected:

VPLEX NDU to GeoSynchrony Releases 6.2 Family 23

Page 29 of 36
The following illustrates explains the workflow when the --io-fwd-ask-for-confirmation on-missing-logins
option is selected:

24 VPLEX NDU to GeoSynchrony Releases 6.2 Family

Page 30 of 36
The following illustrates explains the workflow when the --io-fwd-ask-for-confirmation never option is selected:

VPLEX NDU to GeoSynchrony Releases 6.2 Family 25

Page 31 of 36
Appendix B: Version information
The version details of VPLEX 6.2 family are as follows:
GeoSynchrony 6.2 on VS2:

VPlexcli:/> version -a --verbose

Product Version: 6.2.0.00.00.32

What: SMSv2
Version: 162.0.0.324.0
Build time: January 16, 2020 at 10:48:12AM GMT
Build machine: 1330c0e64ccb
Build OS: Linux version 3.13.0-96-generic on amd64
Build compiler: 1.8.0_222

26 VPLEX NDU to GeoSynchrony Releases 6.2 Family

Page 32 of 36
 Build source: /export/local1/jenkins/clone_legacy/ui

What: Mgmt Server Base

Version: 162.0.0.240

What: Mgmt Server Software

Version: 162.0.0.422

For director /engines/engine-2-1/directors/director-2-1-A:

What: O/S
Version: 162.0.0.240 (SLES11)

What: NSFW
Version: 162.0.0.422.0

What: ZPEM
Version: 162.0.0.422.0-0

What: Director Software

Version: 162.0.0.422.0

What: SSD Model: P30056-MTFDBAA056SAL 118032803

Version: 0005

What: BIOS Rev

Version: 08.50

What: VPLEX Kdriver

Version: 4.1.b_vplex_D40_00_Redwood.3.sh-1

What: ECOM
Version: 162.0.0.232.0-0

What: ZECL
Version: 162.0.0.422.0-0

What: FW Bundle Rev

Version: 12.60

What: POST Rev

Version: 43.80

GeoSynchrony 6.2 on VS6:

VPlexcli:/> version -a --verbose

Product Version: 6.2.0.00.00.32

What: SMSv2
Version: 162.0.0.324.0
Build time: January 16, 2020 at 10:48:12AM GMT
Build machine: 1330c0e64ccb
Build OS: Linux version 3.13.0-96-generic on amd64
Build compiler: 1.8.0_222
Build source: /export/local1/jenkins/clone_legacy/ui

What: Mgmt Server Base

Version: 162.0.0.240

What: Mgmt Server Software

Version: 162.0.0.422

For director /engines/engine-1-1/directors/director-1-1-B:

What: O/S
Version: 162.0.0.240 (SLES11)

VPLEX NDU to GeoSynchrony Releases 6.2 Family 27

Page 33 of 36
 What: NSFW
Version: 162.0.0.422.0

What: ZPEM
Version: 162.0.0.422.0-0

What: SSD Model: SHM2S6D064GHM11EMC 118000106

Version: S8FM08.2

What: Director Software

Version: 162.0.0.422.0

What: BIOS Rev

Version: 41.14

What: BMC Rev

Version: 22.20

What: ECOM
Version: 162.0.0.232.0-0

What: VPLEX Kdriver

Version: 4.1.b_vplex_D40_00_Redwood.3.sh-1

What: ZECL
Version: 162.0.0.422.0-0

What: POST Rev

Version: 31.80

NDU Worksheet
Before beginning your NDU, use this worksheet to record information that you will need during the NDU.

Description Contents
Collect NDU information from the Release Notes and from the NDU Procedure.
From the Release Notes
Upgrade to software version from the Release Notes
(version)
Location of the management server package for the new
GeoSynchrony release (VPlex-version-management-
serverpackage.tar)

Name of the management server package (VPlex-

versionmanagement-server-package.tar)

Checksum of the management server package.

Location of the director software package tar file.
Name of the director software package tar file. (VPlex-
versiondirector-firmware-package.tar)

Checksum of the director software package tar file.

From the Collect Information section of NDU
Upgrade from Release number
Type of system (VPLEX Local, Metro)
Public IP address of management server for cluster 1
Public IP address of management server for cluster 2 (Metro
only)

28 VPLEX NDU to GeoSynchrony Releases 6.2 Family

Page 34 of 36
Description Contents
Service account password
Starting GeoSynchrony version on cluster 1
Does cluster 1 version match cluster 2 version? Yes / No
Will the customer or the storage administrator be available to Yes/No
verify the I/O paths or manually scan the I/O paths on the
hosts during the I/O forwarding phase of the NDU?
Is your VPLEX instance connected to non-production hosts? Yes/No
If your VPLEX instance is connected to non-production hosts, Yes/No/NA
do you have the list of initiators on them?
VPLEX Witness information:
VPLEX Witness server configured (Metro only)? Yes / No
VPLEX Witness public IP address?
For configurations containing a RecoverPoint splitter:
RecoverPoint splitter configured? Yes / No
RPA Host IP address of splitter

VPLEX NDU to GeoSynchrony Releases 6.2 Family 29

Page 35 of 36
Dell Technologies Confidential Information- for internal use only version: 2.9.0.87

Page 36 of 36