PPME FOR HACMP CLUSTERS (AIX HOSTS)?

Hello,

Take a look at the latest EMC® PowerPath® Migration Enabler User Guide.  There is a chapter specifically on this subject.

Migrate in an HACMP environment (AIX hosts)

(from the 5.5 guide)

Note: This section applies only to AIX hosts.

PowerPath Migration Enabler supports migrations in a standby

configuration, where one cluster node is active and the other is

passive.

In the following procedure:

◆ hdiskpower7 = the source device

◆ hdiskpower9 = the target device

Note: Steps 3 through 11 in the following procedure must be done for each

source and target migration pair.

1. Map the intended target device(s) to the active and passive nodes,

and then run the cfgmgr command.

2. Stop the HACMP service on the passive node.

3. On the active node, set up the migration and synchronize the

devices, as described in Chapter 2, ”Migration Procedures.”

4. On the active node, run powermt display dev=all and note the

unique device ID for the target device (hdiskpower9) and the

source device (hdiskpower7) in the migration. The unique device

ID consists of the following information:

• Symmetrix device ID = Symmetrix ID + logical device ID

• VNX/CLARiiON device ID = VNX/CLARiiON ID + logical

device ID (LUN)

5. On the passive node:

a. Run powermt display dev=all, and locate the device whose

ID matches the device ID of the target device, noted in step 4.

(That is, the device ID associated with hdiskpower9 on the

active node.)

b. On the passive node, run chdev -a pv=yes -l hdiskpowerX on

the hdiskpower device identified in step 5 a.

where X is the hdiskpower number identified in step 5.

For example, chdev -a pv=yes -l hdiskpower9

6. Type lspv | grep hdiskpowerX to verify the device you modified

in step 5 b has the same PVID as the source device on the active

node. (Identified in step 4, in this example, the source-device

name is hdiskpower7).

where X is the source device on the active node.

7. Complete the remaining migration steps, as outlined in

Chapter 2, ”Migration Procedures.”

Note: After completing the migration steps, the device names on the

active node only are swapped.

8. On the active node, determine the device ID of the original source

device (in this example, hdiskpower7, which is now the target

device on the active node).

9. On the passive node, run powermt display dev=all, and locate

the device whose device ID is the same as the ID of the original

source device (hdiskpower7, which is now the target device on

the active node).

10. On the passive node, run chdev -a pv=clear -l hdiskpowerX on

the hdiskpower device identified in step 9:

where X is the device identified in step 9.

For example, chdev -a pv=clear -l hdiskpower7

11. Use the lspv command to verify each node has only one device

name associated with the PVID of the target device on the active

node.

12. Synchronize the cluster on both nodes and make sure

synchronization completes.

13. Start HACMP on the passive node, and ensure the cluster service

starts.

14. Fail the resource group from the active to the passive node and

verify the Volume Group is activated.

Thanks for the info .. I already saw in user guide .. they expalin clearly about W2K3 ANS W2K8 , But for HACMP i didnt get the step 5 a, 5b

b. On the passive node, run chdev -a pv=yes -l hdiskpower  ?  This command use after TGT SELECT ?

I had to consult with an AIX expert.  You must add a PVID (physical volume ID, similar to a disk signature in Windows) using the above command.  In Windows, PPME would copy the disk signature from the source to the target disk, modify the signature on the source and unmount it during the powermig cleanup step. 

Thanks .. I got the answer when we need to place PVID  on passive node ..that is after sync complete .. your right we need plave PVID on a device so it can be used by the AIX LVM to identify what disk the volume is ..

Hi Kally,How did you fix it that couldn't get  step 5?   I got "device is busy" when I placed a PVID to the target disk after syncing.  thanks.