PowerHA Implementation

Hands-on Workshop

PowerHA实施专家级课程

© Copyright IBM Corporation 2010

 实施专家级课程 PowerHA

Traditional Client Need: Disaster Recovery

¾ Solution Strategy:
z Recover data & operations from tape
z Separate recovery location
z Primary focus was recovery at a remote location

Data Center Remote Recovery Center

Page 2
 实施专家级课程 PowerHA

Client Need: Multi Site Disaster Recovery

¾ Solution Strategy:
z Manually recover data & operations from replicated
data
z Separate recovery location

z Primary focus is recovery point (RPO)

Remote Site

Data Center

Page 3
 实施专家级课程 PowerHA

Client Need: High Availability – Data Center

¾ Solution Strategy:
z PowerHA SystemMirror Standard Edition
z Near continuous application service
z Eliminate the affects of planned outages
z Minimize unplanned outages
z Regular and sustained role swap operations

ƒ Primary focus is recovery time (RTO) & (RPO)

Page 4
 实施专家级课程 PowerHA

Client Need: Unified HA/DR Solution

¾ Solution Strategy:
ƒ PowerHA SystemMirror Enterprise Edition
ƒ HA clustering for data center & multi site operations

ƒ Focus is a unified solution for data center & multi site resiliency

Remote Site

Data Center

Page 5
 实施专家级课程 PowerHA

PowerHA
High availability solutions for UNIX, i and Linux clients

¾ Cluster management solutions for high availability

z Can enable near continuous application service
z Minimize impact of planned and unplanned outages
¾ PowerHA™ SystemMirror for AIX®
¾ PowerHA for i www.ibm.com/systems/power/software/availability

Page 6
 实施专家级课程 PowerHA

PowerHA SystemMirror Standard and Enterprise Editions

PowerHA SystemMirror for AIX Standard Edition

¾ Cluster management for the data center
z Monitors, detects and reacts to events
z Establishes a heartbeat between the systems
z Enables automatic switch-over

¾ IBM shared storage clustering

z Can enable near-continuous application service
z Minimize impact of planned & unplanned outages
z Ease of use for HA operations

PowerHA SystemMirror for AIX Enterprise Edition

¾ Cluster management for the Enterprise
z Multi site cluster management
z Includes the Standard Edition function
z Optimized for IBM storage
z Supports EMC SRDF

z GLVM(Geographical Logical Volume Manager)

two site configuration wizard

7 Page 7
 实施专家级课程 PowerHA

HACMP监测四类故障
¾ Node Failures
z Processor hardware or operationg system Network
Network
failures Adapter
Failure
Failure
z One or more surviving nodes can acquire
resources
¾ Network Adapter Failures
z Move IP address to standby network adapter in
same node
¾ Network Failure
z Message displayed on console and event is
logged
z As every site's network configurations are
unique , no other default action is taken
z Action to be taken in response to network
failures is customizable
¾ Application Failure
z WebSphere/DB2/Oracle AS &DB
Node
Failure

Application
Failure

Page 8
 实施专家级课程 PowerHA

其它类型的故障
¾ Disk Drive Failures
z LVM Mirroring
z RAID Disk Devices
¾ Other Hardware Failure
z Application Failure (Customization needed , SRC)
¾ HACMP Failure
z Promoted to node failure
¾ Power Failure
z Avoid common power supplies across replicated devices
z Use a UPS

Page 9
 实施专家级课程 PowerHA

Some environment is not suitable for HACMP

¾ You cannot suffer any downtime

z Failovers will cause at least some downtime

¾ Your environment is not stable

z HACMP depends on stable software levels and stable configuration

z HACMP is susceptible to the “fiddle factor”

¾ Your application needs manual intervention to recover from a failure

z Manual reset of a device, etc.

Page 10
 实施专家级课程 PowerHA

HACMP Consideration

¾ Application must be able to recover from a stop/restart operation

z Must release all resources when stopped—either normally or abnormally
z Must tolerate a loss of memory contents
z Must tolerate a loss of processor state
z Must perform a restart from a checkpoint
z Must recover from partial data writes
z Must operate in a “transactional” protocol
¾ There must not be a single point of failure in the HA cluster
z Shared power supply, non-protected disk, etc.
z HACMP is a software solution

Page 11
 实施专家级课程 PowerHA

Some advice for HACMP

¾ Application must be able to recover from a stop/restart operation

z Must release all resources when stopped—either normally or abnormally
z Must tolerate a loss of memory contents
z Must tolerate a loss of processor state
z Must perform a restart from a checkpoint
z Must recover from partial data writes
z Must operate in a “transactional” protocol

¾ There must not be a single point of failure in the HA cluster

z Shared power supply, non-protected disk, etc.
z HACMP is a software solution

Page 12
 实施专家级课程 PowerHA

HACMP软件规划-系统软件
¾ 操作系统的版本和补丁要求
z 信息查看：
http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/TD101347
z 补丁下载
http://www14.software.ibm.com/webapp/set2/sas/f/hacmp/home.html

HACMP Version Supported? AIX Level(s) MISC

5.3 No-9/30/2009 5.2(ML4), 5.3(ML2) or higher AIX 5.2 RSCT 2.3.6 or higher
6.1 AIX 5.3 RSCT 2.4.2 or higher
5.4.0 Yes 5.2 (TL8) or higher AIX 5.2 RSCT 2.3.9 or higher
and 5.3(TL4) or higher AIX 5.3 RSCT 2.4.5. or higher
5.4.1
5.4.1 Yes 6.1

5.5 Yes 5.3 (TL9 ) or higher AIX 5.3 RSCT 2.4.10.0 or higher
6.1 (TL2,SP1) AIX 6.1 RSCT 2.5.2.0 or higher
release_notes
6.1 Yes AIX 5.3 TL 9 AIX 5.3 RSCT 2.4.12.0 or higher
AIX 6.1 TL 2 AIX 6.1 RSCT 2.5.4.0 or higher
release_notes

Page 13
 实施专家级课程 PowerHA

PowerHA SystemMirror C-SPOC Tools

¾ Cluster-Single Point of Control, the System Management interface in

HACMP
¾ Cluster wide, cluster aware tool
z Add, change, delete users

z Add, change, delete file systems

z Add, change, delete logical volumes

z Add, change, delete physical volumes

z Start and stop the cluster

z Manage log files

z File collections

z Manage Resource Group State

z …

Page 14
 实施专家级课程 PowerHA

PowerHA SystemMirror with WebSMIT

¾ Observe status of all your clusters

z Get the status of a cluster
z Get the status of a node within a cluster
z Generate reports on any cluster as desired
¾ Manage a cluster
ƒ Make configuration changes
z Add, remove, edit, create HACMP
configuration
¾ Install WebSMIT on one system
z No footprint of WebSMIT on managed nodes
z No performance impact on managed clusters
z Single sign-on for all of your WebSMIT
managed clusters
¾ Security management
z Implement WebSMIT on a system outside of
the managed clusters
z Same communication daemon as used by
HACMP (clcomd) with the same encryption
technology
z Access control for all of your WebSmit users
z Enable administrators with read only access
Page 15
 实施专家级课程 PowerHA

参考资料
¾ 1. PowerHA for AIX Cookbook
http://www.redbooks.ibm.com/abstracts/sg247739.html
¾ PowerHA Website
z www.ibm.com/systems/power/software/availability/

¾ IBM InfoCenter
z http://publib.boulder.ibm.com/infocenter/systems/scope/aix/index.jsp

ƒ HACMP for AIX: Concepts and Facilities Guide SC23-4864

ƒ HACMP for AIX: Planning Guide SC23-4861
ƒ HACMP for AIX: Installation Guide SC23-5209
ƒ HACMP for AIX: Administration Guide SC23-4862
ƒ HACMP for AIX: Troubleshooting Guide SC23-5177
ƒ HACMP for AIX: Programming Client Applications SC23-4865
ƒ HACMP for AIX: Master Glossary SC23-4867

ƒ HACMP/XD GLVM Planning and Administration Guide SA23-1338

ƒ HACMP/XD: Metro Mirror Planning and Administration Guide SC23-4863

ƒ HACMP for AIX: Smart Assist for WebSphere User's Guide SC23-4877
ƒ HACMP for AIX: Smart Assist for Oracle SC23-5178
ƒ HACMP for AIX: Smart Assist for DB2 SC23-5179
ƒ HACMP for AIX: Smart Assist Developer's Guide SC23-5210

Page 16
Thank
You!
© Copyright IBM Corporation 2010