Transportable tablespaces in 9.7.2
in [DB2]
|
Prev: DB2 V10 and z800
Next: Question on DB2MON
From: hsn_ on 9 Jun 2010 03:19 Problems with transportable tablespaces was due to COMMENTs on objects being transported.
From: Helmut Tessarek on 9 Jun 2010 12:58 > Its CSV export from procmon. procmon is something like systrace for > windows. It records kernel calls made by application. Ok, looked a little bit more detailed than the Event Log anyway. :-) Unfortunately I can't see what is going on in the DB2 code from this trace. I would need a db2trc dump, formatted as format and flow trace. > I am looking into our bug database and one customer had this problem > too, diaglevel 4 on windows and db2 does not start anymore. > Its timing dependant race condition - most likely 2 threads in > db2starts fights against lock on diag file. It appears on diaglevel 4 > more because its needed to grab diag lock sooner due to more extensive > log output. According to your procmon output, the locking issue has something to do with rotating the logs. It does not seem to be a problem with db2diag.log itself. If you change DIAGLEVEL to 3 again, db2start works? If you set it then to 4, db2start fails again? How does it fail? Does it hang? If yes, a stack trace would be nice. db2pd -stack all This is a problem that should be handled via a PMR. If you have a PMR, please let me know the number. -- Helmut K. C. Tessarek DB2 Performance and Development /* Thou shalt not follow the NULL pointer for chaos and madness await thee at its end. */
From: hsn_ on 9 Jun 2010 16:40 I created new package for you. http://rapidshare.com/files/397201033/package.zip it contrains db2trace output and procmon syscall trace dump. You need procmon from http://technet.microsoft.com/cs-cz/sysinternals/bb896645%28en-us%29.aspx to view it. db2pd -stack -all hangs producing no output. > If you change DIAGLEVEL to 3 again, db2start works? yes > If you set it then to 4, db2start fails again? no > How does it fail? Does it hang? yes i think i found way to re-create this issue. set diaglevel to 4 then db2stop, then delete .db2diag.rotate.lck and db2diag.*.log db2start hangs. let me know if it worked for you too. i am not aware that we have opened some pmr for it, we adopted dont run db2 with diaglevel 4 on windows as part of our best practices
From: Helmut Tessarek on 9 Jun 2010 17:38
> I created new package for you. > http://rapidshare.com/files/397201033/package.zip Thank you for the data. > db2pd -stack -all hangs producing no output. Hmm, this should not happen either. Very odd. >> If you change DIAGLEVEL to 3 again, db2start works? > yes At least it is not a permanent error. :-) >> If you set it then to 4, db2start fails again? > no Now, I'm confused. When does db2start fail? Only, if you set DIAGLEVEL to 4 and delete the .db2diag.rotate.lck and db2diag.*.log files? It does not fail, if you don't delete the files? So you can actually set it to 4 and start db2 somehow? > i think i found way to re-create this issue. set diaglevel to 4 then > db2stop, then delete .db2diag.rotate.lck and db2diag.*.log > db2start hangs. let me know if it worked for you too. I tried it, but I'm still not able to reproduce it. But I tried it on Win32. What is your 'db2level' output? Can you please also post the output of 'db2 get dbm cfg'? As soon as I'm able to reproduce the problem, I'll send the data and the problem description to the owner of this component. > i am not aware that we have opened some pmr for it, we adopted dont > run db2 with diaglevel 4 on windows as part of our best practices Usually DIAGLEVEL 3 is more than enough, but for deeper problem analysis it makes sense to change it to 4. -- Helmut K. C. Tessarek DB2 Performance and Development /* Thou shalt not follow the NULL pointer for chaos and madness await thee at its end. */ |