As part of my resolution of server failure over the weekend I ran a Recovery Agent extract on the target replica. In the end we decided not to manually apply any of the AIs due to there not being much work lost as a result of not applying it. But for future reference, I was wondering if someone could answer me something.
Just for clarity - the command I issued was: $ dsrutil target -C RECOVERY agent > recagent.out
Here is that file:
Online Replication Recovery Information for F:\DATABASE\LIVE\icmasliv:
Replication version: 5.0
Date created: Mon Apr 21 12:43:06 2014
Date last written: Thu Apr 24 18:02:25 2014
Replication local agent information:
Last Block: Incomplete
Last block received location: area: 87, seq: 7687, loc: 134152192, offset: 0
Last block processed location: area: 0, seq: 0, loc: 0, offset: 0
Last block ACKed location: area: 87, seq: 7687, loc: 133685248, offset: 4187
Last block received: no date
Last block ACKed: no date
ID of the last TX begin: 1731109729
ID of the last TX end: 1731109730
Time of last TX end: Thu Apr 24 18:00:43 2014
Last AI Extent processed
AIMAGE BEGIN date: Sat May 25 02:16:06 2013
AIMAGE NEW date: Thu Apr 24 17:01:54 2014
After-Image File Number: 7687
File Last Opened: Thu Apr 24 17:01:54 2014
Completely Applied to Target: No
The docs I was looking at say that the AI File Number is relevant. Except we don't have 7687 files. Now I notice that the Last block received location is area 87. And 87 is the last 2 digits of the file number. Coincidence?
Essentially, how would I have decided which AI(s) needed to be applied?
The 7687 is the sequence number, not the extent number. If you are using the AIFMD, then when the AI file is archived out, the sequence number is embedded in the archived AI file name. If you run "rfutil DB -C aimage list" on your PROD DB, you;ll see the sequence number of all the non-EMPTY AI extents.
Thanks again Paul :)