Recall process interrupted with "Severing connection" error

[tivolimistery]

Dear All,

I'm trying to recall some old data from our very old TSM Server v.5.5.2 before decomissioning it.
Data recalls work as expected for a lot of time but sometimes recalls processes are interrupted suddenly with the following error ">>> Process interrupted...severing connection <<<".

I checked the client logs, but I didn't find any useful information. If I try to restart recall processes, they go ahead for a while and, in some cases, they stop again with the same error.

Could you please help me to understand what is happening and where?

Thank you very much,
Mauro

 

[LED888]

TSM Server 5.5.2, what is is the os and the full version of the os that is hosting the TSM Server?
What is the full version of the TSM B/A Client / HSM and os?


What are the messages in the actlog around the time frame of when the error message was produce/
Can you post, or upload, the actlog and the dsmerror.log?

Good Luck,
Sias

 

[tivolimistery]

Hello Sias,

you are right, I should add some additional info.

TSM Server OS: AIX 5.3
TSM server: 5.5.2

TSM Client OS: CentOS 6.2 Final
TSM B/A Client: 6.4.0
HSM Client: 6.4.0

No relevant or related messages in actlog.
dsmerror.log on TSM server is empty.
In attachment you can find the client dsmerror.log .

Many thanks for your availability.
I really appreciated it.
Mauro

 

[LED888]

From the dsmerror.log :

03/22/2021 18:38:20 ANS9507E dsmreconcile: cannot request the right on session 6058D39C00000000 for file handle 0000000010ACF001-B41A80730323F905-6C7B077AAF41CB3C-0000000000000000
token = A-6058D60C000000AE. Reason : Invalid argument

03/22/2021 18:38:20 ANS9616E dsmreconcile: cannot get migration information for /cosmomed/casarchive/rcp45/rcp45_cas201404.tar on /archive

03/22/2021 19:10:34 ANS9101E Migrated files matching '/archive/ans033/EXPERIMENTS/PELAGOS/DATA/PELAGOS025OFF-MYO-V5/outdata/bfm/PELAGOS025OFF_1m_20011201_20011231_bfm.tar.gz' could not be found.


From my black book, it appears that I encounter this many moons ago.
Sadly my documentation was not up to par, I mention the techdoc
ANS9616E dsmreconcile: cannot get migration information for /fs/file1
No indication if that resolve my issue, because co-worker upgraded the B/A Client and HSM to be within supported levels.

I my doc, there was a known issue, IT04406 Reconcile fixes corrupt file handles.
Of course my weak documentation at the time mention that the issue is address at HSM 6.4.3 .

TSM for Space Management (HSM) APARs fixed in 6.4.3
If there is a potential of data loss, would expect to see a FLASH on this.

NOTE: The TSM B/A Client 6.4 is not compatible with TSM Server 5.5.
Technically it should work but the results are unknown.

Chat with a co-worker, and he say it could be an indication of a general file system corruption


Good Luck,
Sias

 

[tivolimistery]

Hello Sias,

thank you very much for your help.
So, if I understand, we can say that:

- TSM B/A client is not compatible with TSM server 5.5;
- Also HSM client is not compatible with the server version;
- dsmreconcile is trying to fixes some corrupted file handles;
- it could be an indication of a general file system corruption.

Is it correct?
Many thanks.
Have a great day,
Mauro

 

[LED888]

Would say yes, unless someone else have encounter the same issue with the recall.

As for the "general file system corruption" not sure how my co-worker came up with this conclusion.

Good Luck,
Sias