Thanks. This is exactly what I needed. We altered it to used mail and send the info to me. I want to look at it before canceling the session., so we can talk to the user. Zoltan Forray TSM Software & Hardware Administrator Virginia Commonwealth University UCC/Office of Technology Services [email protected] - 804-828-4807 Don't be a phishing victim - VCU and other reputable organizations will never use email to request that you reply with your password, social security number or confidential personal information. For more details visit http://infosecurity.vcu.edu/phishing.html
From: Jean-Paul Duyx <[email protected]> To: [email protected] Date: 04/08/2010 10:31 AM Subject: Re: [ADSM-L] Log issues and scheduled commands and CHANGINGRETRIES Sent by: "ADSM: Dist Stor Manager" <[email protected]> Hi Zoltan, We have similar issues on one of our TSM instances that has some clients behind a slower connection. We cancel the sessions that pins the logs. And we are in a process of moving some nodes away from this instance, to another instance. We have put a script in CRON that checks every 15 minutes and cancels the session that is pinning the log when the log is over 55 percent full. We want to keep our TSM server alive, and are forced to cancel sessions that keep the log pinned. Maybe canceling at a higher percentage filled log could work for you. {{{ #!/usr/bin/sh PROCENT=`/usr/bin/dsmadmc -se=cestino -id=admin -pa=<****> -dataonly=y "select '##',cast(PCT_UTILIZED as int) from log" |grep "##"|grep -v grep| awk '{ print $2 }'` if [ $PROCENT -gt 55 ]; then echo "Recovery Log is $PROCENT procent" echo "Issuing show logpinned cancel" /usr/bin/dsmadmc -se=cestino -id=admin -pa=<****> -dataonly=y show logpinned cancel Fi }}} Line that starts with PERCENT= Should be on one line, but I am afraid mailers will break it. Good luck ! Jean-Paul Duyx -----Original Message----- From: ADSM: Dist Stor Manager [mailto:[email protected]] On Behalf Of Zoltan Forray/AC/VCU Sent: donderdag 8 april 2010 16:17 To: [email protected] Subject: [ADSM-L] Log issues and scheduled commands and CHANGINGRETRIES Ever since I upgrade one of our busy TSM servers from 5.5.3 to 5.5.4.1, I have been having constant problems with the log being pinned and filling. While I was able to determine and fix 2 problem nodes (long/large Oracle backup and another node that is compressing and backing up a 280GB log file, 5-times and thus running for 10-hours), I am still having problems. The log went to 90% last night. I see there is a 5.5.4.2 server patch but none of the fixes I saw addressed such issues. So, I would like to schedule a "show logpinned" command every 15-minutes and send the output somewhere so it can be emailed to me. I have never messed with server scripting so this is new territory for me. Does anyone have a script/process they can share, to deal with this? The TSM servers are RedHat Linux 5. One a related topic, I am constantly finding folks who have live/changing files being backed up multiple times (yes, I know - server owners should be more responsible for what they are backing up....), since the default is "CHANGINGRETRIES 5". I am seriously thinking of pushing "CHANGINGRETRIES 0" via CLOPTSET. Is there and real value to the default of "5"? What do folks out there set this to and why? Zoltan Forray TSM Software & Hardware Administrator Virginia Commonwealth University UCC/Office of Technology Services [email protected] - 804-828-4807 Don't be a phishing victim - VCU and other reputable organizations will never use email to request that you reply with your password, social security number or confidential personal information. For more details visit http://infosecurity.vcu.edu/phishing.html
