We have had a situation where we had a number of rapid after-image switches that were caused by user activity. We *believe* that when this occurrs, the user whose activity is forcing the switch is noted in the .lg file where the log will say:
Usr 23: (3777) Switched to ai extent <file>
Usr 23: (3778) This is after-image file number <number> since the last AIMAGE BEGIN
In other words - user 23 above would be the user whose activity forced the AI switch.
Is this true? We're trying to find out the source of a what looks to have been a runaway transaction.
oe10.1b SP3/ HPUX 11.11
The user that wrote those messages is the one that triggered the extent switch. It did so because it had an after-image record to write and had to do the swtich before it could write the data. That does not necessarily mean it was the same user that filled up the after-image extent though that is somewhat likely. It could have been a different user though.
If you make a copy of that extent (the one that got filled) and then run an aimage scan verbose on it, you will be bale to see which user wrote the most stuff to it. In the output, you will see lines that look like this (but with other users lines mixed in):
Trid: 18 User Id: Gus (12531)
Trid: 18 code = RL_TBGN version = 1 (12528)
Trid: 18 dbkey = 0 update counter = 0 (12530)
Trid: 18 code = RL_RMCR version = 2 (12528)
Trid: 18 area = 7 dbkey = 480 update counter = 5 (12529)
Trid: 18 code = RL_CXINS version = 2 (12528)
The trid is the transaction id, 18 in this case, started by user Gus and all the records with trid 18 were written as part of that same transaction by the same user.
This is exactly what I needed - Thanks, Gus!!