01-21-2009 01:35 PM
I have a vault job that fails every day (m-f), this job is allowed to use two tape drives for writing and one for reading.
The primary vault job kicks off at lets say 4:00 am it then allocates 3 tapes for the duplication part.
Ok the Vault job start at 4 am everything is good, images are moving and shaking to the tapes, two hours later the vault job fails with a 288, a 288 means that one or more of the three tapes it allocated above was still busy writing data to the tape.
Now why the heck does the vault job continue to run then stop before the images are written to the tape?
At this point a tape is still being written to. Gee but vault thinks they are ready to eject.
04:29:18.638 [24299] <2> vltrun@VaultRobot::pruneMediaBasedOnDbInfo^2254: Found 3 media in the volume database to eject
Loading them into the eject.list ;)
04:29:20.654 [24299] <4> vltrun@VaultSession::getMediaForEject()^2254: Loaded 3 media from eject.list file
Oh yeah master server wants to suspend them first.
04:29:22.999 [24299] <2> vltrun@VaultRobot::suspendMediaToEject()^2254: For media 007033, suspend host is master_server
04:29:23.207 [24299] <2> vltrun@SuspendMedia^2254: media 007033 with host is suspended successfully
Works great for the fisrt two tapes because they are not in a tape drive they are in their cozy slots waiting to be ejected.
But wait here comes the last one, it still in a tape drive doing whats its supose to do (get images written to it)
04:29:23.369 [24299] <16> vltrun@SuspendMedia^2254: Suspend FAILed MEDIA=007126, HOST=master. EMM Err=2001049 ( Media is already allocated )
04:29:23.370 [24299] <16> vltrun@SuspendMedia^2254 FAILed MM_EC=199 MM_MSG=the media is allocated for use
04:29:23.377 [24299] <16> vltrun@SuspendMedia^2254: Leaving with DMN=2 SC=199
04:29:23.377 [24299] <16> vltrun@VaultRobot::suspendMediaToEject()^2254: suspend failed for media 007126, ignoring...
04:29:23.377 [24299] <16> vltrun@VaultRobot::suspendMediaToEject()^2254 FAILed NB_EC=97 NB_MSG=requested media id is in use, cannot process request
Duh it cannot prosses the request because its still trying to write to the tape it allocated for the vault proccess.
So why is the vault job getting ahead of itself?
Jim