Tuesday, 2010-12-07

*** stliu has joined #ingres00:05
*** grantc has joined #ingres01:06
*** PaulM05 has joined #ingres01:08
*** stliu has quit IRC02:17
sunny_slougha sysmod on a prod db failed today with "E_QE0024_TRANSACTION_ABORTED The transaction log file is full" while processing iiattribute. we have a 2gb tx log file and iiattribute has 2160610 rows. What should I look for to identify the cause?03:09
PaulM05what else was running?03:10
sunny_sloughthis was at 06:33, overnight batch completed at  01:00, backups completed at 05:30, no users logged in...03:11
PaulM05no other sessions? no other batch jobs? possibly on other databases?03:12
sunny_sloughno other sessions,  all batch work complete, no other database active...03:14
PaulM05was the archiver running?03:15
sunny_sloughyes the archiver was active.03:17
PaulM05hard to say without seeing the state of the logfile at the time. You're either looking from something holding open the BOF, or the sysmod itself can't fit in the logfile - which seems unlikely03:20
PaulM05any errors in the logs around that time?03:22
sunny_sloughsysmod issued E_US1262 Your transaction has been externally aborted.e rrlog contains E_DM010C_TRAN_ABORTED, E_QE0024_TRANSACTION_ABORTED, E_DM9059_TRAN_FORCE_ABORT...03:24
sunny_sloughwhat is 'the BOF'?03:25
PaulM05Beginning Of File - the logical start of the logfile, it gets moved forward when committed transactions are flushed by the archiver, if it doesn't move then the log will eventually fill up03:26
PaulM05typical reasons are a session not committing, or archiver not running (therefore not flushing records)03:27
sunny_sloughthe sysmod that failed was the first of four, the remaining three completed ok, they were followed by successful checkpoint of iidbdb...03:30
PaulM05so probably not that then - must be that the sysmod itself took a lot logging for some reason. Are the other system catalogs quite large too?03:31
sunny_sloughthe other system catalogs are small..03:36
PaulM05what version of Ingres is this?03:37
sunny_sloughrunning 9.1.1 on solaris1003:38
PaulM05I guess the other question is what has changed since the last time it ran successfully?03:47
sunny_sloughIt last ran successfully on Friday,3 dec... we did have a problem with the server yesterday, a sysmod on iidbdb took over 7 hours to complete.03:54
sunny_sloughWe decided to reboot the server after the iidbd sysmod finished because a test to retrieve data from the online primary application failed (did not respond)03:55
PaulM05do you have support? If so I'd log an issue03:56
sunny_sloughWe have support via Capita - the application provider, when we reported yesterday's problem they said we should not run sysmod on iidbdb, is that good advice?04:00
sunny_sloughI will run today's problem past them too....04:03
PaulM05sorry - phone04:04
PaulM05re: sysmod on iidbdb - I wouldn't say it's necessarily always a bad idea but it will lock out iidbdb which could be a problem, also it's unlikely to have much of a benefit on most systems04:06
sunny_sloughthank you for generously giving your time this morning, issue logged with application provider....04:19
PaulM05no problem04:21
*** atrofast has joined #ingres04:32
*** atrofast has joined #ingres04:33
*** cthibert has joined #ingres04:43
*** stliu has joined #ingres05:04
*** stliu has joined #ingres05:04
*** KermitTheFragger has joined #ingres05:19
*** mull has quit IRC05:38
*** mull has joined #ingres06:19
*** stliu has quit IRC06:46
*** Alex| has quit IRC07:21
*** Alex| has joined #ingres08:33
*** ChanServ sets mode: +o Alex|08:33
*** Alex| has quit IRC09:32
*** cthibert has left #ingres09:37
*** cthibert has joined #ingres09:39
*** DerMeister has joined #ingres10:47
*** PaulM05 has quit IRC10:59
*** weta has quit IRC11:08
*** Mud has quit IRC11:14
*** weta has joined #ingres11:17
*** DerMeister has quit IRC11:40
*** KermitTheFragger has quit IRC13:03
*** cthibert has left #ingres13:10
*** mull has quit IRC14:23
*** grantc has quit IRC15:27
*** sunny_slough has quit IRC17:36
*** stliu has joined #ingres17:56
*** stliu has quit IRC18:01
*** stliu has joined #ingres18:02
*** stliu has joined #ingres18:02
*** atrofast has quit IRC18:04
*** cytrinox_ has joined #ingres21:08
*** cytrinox has quit IRC21:11
*** cytrinox_ is now known as cytrinox21:11
*** Alex| has joined #ingres22:38
*** ChanServ sets mode: +o Alex|22:38
*** grantc has joined #ingres23:03
*** Alex| has quit IRC23:23
*** Alex| has joined #ingres23:48
*** ChanServ sets mode: +o Alex|23:48

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!