Log pinned

[oguinna]

I'm relatively new to TSM so please be merciful. We are consistently having to halt and restart TSM on an AIX server platform due to the log being pinned. I’ve got a script in place to show logpinned when the 80 or 90% threshold is breached and it shows the same Windows cluster each time.

The file that appears to be pinning the log is 200 GB.

*** OPT file
passwordaccess generate
commmethod TCPIP
schedmode prompted
tcpport xxxxxx
tcpserveraddress xxxxxxxxxx
tcpbuffsize 32
tcpwindowsize 63
tcpnodelay yes
largecommbuffers yes
txnbytelimit 25600
maxcmdretries 2
changingretries 2
ERRORLOGNAME "C:\Program Files\Tivoli\TSM\baclient\dsmerror_clust.log"
errorlogretention 10 D
SCHEDLOGNAME "C:\Program Files\Tivoli\TSM\baclient\dsmsched_clust.log"
schedlogretention 10 D
quiet
MANAGEDSERVICES WEBCLIENT
HTTPPORT xxxx
nodename xxxxx
clusternode yes

* Excludes for SQL servers
Exclude.File "*:\...\*.mdf"
Exclude.File "*:\...\*.ldf"
Exclude.File "*:\...\*.ndf"
Exclude.File "*:\...\cache.dat"

show logpinned

Thu May 26 06:40:02 PDT 2011 Log % - 91.7
Thu May 26 06:40:02 PDT 2011 TSM log space utilization is 91.7
Thu May 26 06:40:02 PDT 2011 IBM Tivoli Storage Manager
Command Line Administrative Interface - Version 5, Release 5, Level 1.0
(c) Copyright by IBM Corporation and other(s) 1990, 2008. All Rights Reserved.
Session established with server xxxxxx: AIX-RS/6000
Server Version 5, Release 5, Level 4.2
Server date/time: 05/26/11 06:40:02 Last access: 05/26/11 06:20:01
ANS8000I Server command: 'show logpinned'
Dirty page Lsn=12503804.34.3356, Last DB backup Lsn=12503804.93.3140, Transaction table Lsn=12491941.105.819, Running DB backup Lsn=0.0.0, Log truncation Lsn=12491941.105.819
Lsn=12491941.105.819, Owner=DB, Length=64
Type=Update, Flags=C2, Action=SetRange, Page=1036907, Tsn=0:1.3318249890, PrevLsn=0.0.0, UndoNextLsn=0.0.0, UpdtLsn=12491938.246.1791 ===> Start=5183, Count=64
(29247) Generating SM Context Report:
(29247) *** no sessions found ***
(212) Generating SM Context Report:
(212) *** no sessions found ***
(29239) Generating SM Context Report:
(29239) Session 20636: Type=Node, Id=Servername
(29239) Platform=WinNT, NodeId=1299, Owner=
(29239) SessType=4, Index=7, TermReason=4
(29239) RecvWaitTime=0.000 (samples=0)
(29239) Backup Objects ( bytes ) Inserted: 48 ( 0.330239394 )
(29239) Backup Objects ( bytes ) Restored: 0 ( 0.0 )
(29239) Archive Objects ( bytes ) Inserted: 0 ( 0.0 )
(29239) Archive Objects ( bytes ) Retrieved: 0 ( 0.0 )
(29239) Last Verb ( ConfirmResp ), Last Verb State ( Sent )
Session 20636 is assoicated with this transaction.

 

[oguinna]

Since it's a cluster in prod I won't be able to update the client w/o going through proper channels. I've renamed the node and started a backup on the directory I believe to be causing the issue.

Also found both nodes and cluster were backing up on same sched. I suspect this file might be locked and tsm is attempting to resend but I'm not sure.

 

[oguinna]

Turns out it was more of a client side issue. I created a new node on another TSM server to see if we would get better results but we were seeing resource utilization errors on the client side. The cluster was failed over and each node was rebooted. Since then all is well. Thanks