ANS2050E Error - Problem with MS cluster password

[primags]

I am trying to get a TSM node setup to be able to fail over between two MS cluster servers.

I have this working in several other instances, but for some reason this one node is giving me issues.

Here is what I do:
1. In TSM Manager, I find the node and reset the password to "1234".
2. I go to the active server and run DSMC query session -optfile=o:\tivoli\dsm.opt
3. It prompts me for node. I hit enter. It prompts me for password, I put in 1234. It accepts password. I bring the resource online in cluster manager and all is good.
4. I fail over to what was the passive node. I run the same DSMC query session -optfile=o:\tivoli\dsm.opt command and receive the error: ANSI12621 Password is not udpated. Either an invalid current password was supplied or the new password does not fulfill the server password requirements". At this point the service will not start.
5. If I fail back to the other node, it also will not start.

It seems as if when the fail over occurs, the password is no longer valid. I can get TSM working on either node just fine if I change the password back to "1234" and then start the service on whatever node is active. but as soon as failover occurs, the other node acts like the 1234 password is not the correct password.

Here is a copy of the opt file that is on each cluster node

NODENAME %server%
COMMmethod TCPip
TCPPort 1504
TCPServeraddress %tsmserver%
PASSWORDACCESS GENERATE
ERRORLOGRETENTION 5,D
SCHEDLOGRETENTION 5,D
MANAGEDSERVICES WEBCLIENT
HTTPPORT 1581
SNAPSHOTCACHESIZE 5
Clusternode Yes
DOMAIN O:
DOMAIN P:
errorlogname o:\tivoli\dsmerror.log
schedlogname o:\tivoli\schedlog.log



Any ideas?

 

[moon-buddy]

The problem is the password for the Cluster that was set or defined when you created the shared resource using the MS Cluster Manager. This is in the registry.

1. Delete the shared TSM service that you defined using the MS Cluster from both nodes,
2. Set or reset the virtual (Cluster nodename) node's password from within the TSM SERVER (more: from WITHIN THE TSM SERVER command line or ISC), and
3. Recreate the shared TSM service using the password you just set or reset. Pay attention to the registry entry when you create the service using the MS Cluster Manager.

Follow the install instruction that is in the Admin Guide for Windows. I believe this is in Appendix B.

 

[primags]

Thanks moon-buddy for your reply.

This was working previously and is not now, with nothing changed in cluster manager.

Have you experienced this before, and does the fix below still apply?

Thanks
Glenn

 

[kidquick10]

Reseting MSCS Cluster Service

Primags-

What Moon-Buddy is saying is Correct.

In addition these steps have worked for me in the past:

1. Delete password entry from the registry (if present) (HKLM\SOFTWARE\IBM\ADSM\CurrentVersion\BackupClient\Nodes\<nodename>\
2. Uninstall the TSM Scheduler service (from the client)
3. Update the password for that node on the TSM Server
4. Update the password via the command line on the client (dsmc)
5. Reinstall the TSM Scheduler service (using the "exact" name) used on the other clustered nodes
6. Restart the service

TSM in a MSCS environment is fun ISN'T IT?

kidquick10

 

[hy92]

Thanks guys. I had exact same problem as primags. Moon-buddy's suggestion worked like charm.

I'm still scratching my my head why it so suddenly could not authenticate. There's no password expiration set for admin or node.