How To
Summary
This document addresses the needed steps on how to migrate PowerHA® SystemMirror® cluster nodes from an older storage unit to a newer one.
Objective
The general framework to migrate storage for PowerHA® SystemMirror® cluster nodes can be summarized as follows:
- Add, and discover the new disks to the cluster nodes.
- Add the volumes to the cluster shared volume group.
- Mirror the data from the old set of disks to the newly added disks.
- Unmirror the data off the old set of disks.
- Remove the old disks from all cluster nodes.
- Follow the hereunder described method to migrate a CAA (Cluster Aware AIX®) repository disk.
Environment
PowerHA® SystemMirror® for AIX version 7.2 and higher.
Steps
A. Add the new disks physically and get them recognized by the system:
1- Add the disk physically
2- Use the following commands to discover the new disks:
# cfgmgr => on all nodes
# chdev -l hdiskX -a pv=yes => on all disks on all nodes.
B. Add the disks to the volume group:
# smit cspoc
> Storage
> Volume Groups
> Set Characteristics of a Volume Group
> Add a Volume to a Volume Group
> Storage
> Volume Groups
> Set Characteristics of a Volume Group
> Add a Volume to a Volume Group
Choose the volume group in question and hit Enter. Then, it should give you a list of available disks. Choose the new ones.
C. Mirror the new disk in the VG:
# smit cspoc
> Storage
> Volume Groups
> Mirror a Volume Group
> Storage
> Volume Groups
> Mirror a Volume Group
NOTE: With the option"Mirror sync mode" set as "Background", the sync runs in the background. In a scenario where a client is having a large amount of data, this operation might take a long time to complete. Also, many of the basic commands like lsvg -l vg will fail to run.
D. Unmirror the old disks from the VG:
# smit cspoc
> Storage
> Volume Groups
> Unmirror a Volume Group
> Storage
> Volume Groups
> Unmirror a Volume Group
E. Remove the old disks from the VG:
Note: Removing a volume using CSPOC only allows for one hdisk to removed per operation. If you are scripting this process, a cspoc command must be issued for each hdisk.
# smit cspoc
> Storage
> Volume Groups
> Set Characteristics of a Volume Group
> Remove a Volume from a Volume Group
> Storage
> Volume Groups
> Set Characteristics of a Volume Group
> Remove a Volume from a Volume Group
F. Remove the old disks from the cluster:
Note: Removing a volume using CSPOC only allows for one hdisk to removed per operation. If you are scripting this process, a cspoc command must be issued for each hdisk.
# smitty cspoc
> Storage
> Physical Volumes
> Remove a Disk From the Cluster
> Storage
> Physical Volumes
> Remove a Disk From the Cluster
NOTE:
The above methods can't be used for replacing the repository disk. To replace the repository disk, you can go through the following steps:
1. Ensure that you have a new shared disk that is accessible by all cluster nodes. It must have a PVID that is known to each node in the cluster.
2. Add that disk as a repository disk by using:
# smit sysmirror
# smit sysmirror
> Cluster Nodes and Networks
> Manage Repository Disks
> Add a Repository Disk.
Choose the disk either through the F4 list (picklist) or manually.
Choose the disk either through the F4 list (picklist) or manually.
3. Run the following command:
# smitty sysmirror
> Problem Determination Tools
> Replace the primary Repository Disk.
4. If sites are defined, you can select a site through the picklist. Otherwise, you are directed to the last SMIT menu.
5. Select the new repository disk by pressing F4.
6. Synchronize the cluster.
Additional Information
In case if a complete downtime is approved, the following method also can be followed for an offline migration:
1- Stop PowerHA and bring the resource groups offline so that volume groups go to varied off state by using the smit clstop command.
2- Run the rmdev command for the LUNs from both nodes (rmdev -dl hdisk#).
3- Remap the SAN storage access to the new storage where the LUN copies have been created.
4- Run the cfgmgr command on all PowerHA nodes to find the new LUNs so they will be discovered as the new hdisks.
5- Start the PowerHA cluster services on both nodes again (smit clstart).
2- Run the rmdev command for the LUNs from both nodes (rmdev -dl hdisk#).
3- Remap the SAN storage access to the new storage where the LUN copies have been created.
4- Run the cfgmgr command on all PowerHA nodes to find the new LUNs so they will be discovered as the new hdisks.
5- Start the PowerHA cluster services on both nodes again (smit clstart).
Also, for the repository disk, the previously illustrated steps above should be followed.
Document Location
Worldwide
[{"Type":"MASTER","Line of Business":{"code":"LOB57","label":"Power"},"Business Unit":{"code":"BU058","label":"IBM Infrastructure w\/TPS"},"Product":{"code":"SSLM9V","label":"PowerHA SystemMirror Standard Edition for AIX"},"ARM Category":[{"code":"a8m3p000000hAvBAAU","label":"PowerHA System Mirror->Administrating PowerHA"}],"ARM Case Number":"","Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"7.1.e;and future releases"}]
Was this topic helpful?
Document Information
Modified date:
30 November 2021
UID
ibm16467725