Replaying Log Files

[royalmail]

I am currently messing around with a test Exchange environment, and was wondering how to replay recent log files. I am trying the following:

Stop services and back up all Exchange data to a different folder.

Restart services, and send a few post backup mails through the system.

Stop services, delete the live versions of dir.edb, priv.edb, and pub.edb.

Copy the backup databases from the backup folder to their original location.

Everything starts fine, but I have lost the the emails that I sent after I made the backup of the databases. As I still have the most recent log files, why hasn't Exchange replayed them so that the database is up to data again?

 

[Alshrim]

Delete the Checkpoint files. Alshrim
System Administrator
MCSE, MCP+Internet

 

[royalmail]

I tried moving the checkpoint files out and restarting the services but then the information store won't start and I get 4294966781 service specific error. If I move the checkpoint files back in and restart the services all is well but I have still lost the mails that I am trying to recover.

 

[smoss51]

In step one - did you backup all edb log files too? There are some Q articles about restoring/replaying log files:

Check 181798, 262041, 296787, 275602, 328639

These may or may not apply to you situation, but they'll give you a pretty good idea what is expected to re-play log files.

 

[royalmail]

Many thanks. I got it all working nicely in the end with a few more attempts. For what it is worth, I put this procedure together for a disaster recovery document I am working on. I haven't tried this with an online backup yet, but it works well for an offline backup with circular logging disabled.

1 Stop all Exchange services apart from the System Attendant and make an offline backup of everything in the exchsrvr\mdbdata folder. If possible, just copy the data out into a backup folder on the same server. If size limitations mean this isn’t possible, just copy all edb*****.log files and the edb.chk file.

2 If there is enough space, rename the corrupt priv.edb/pub.edb file just so that there is a backup of the database in its current state. If space is an issue, make a backup of this file and then delete it.

3 Restore the functional priv.edb/pub.edb database file from backup into the exchsrvr\mdbdata folder.

4 Delete the edb.chk file from the exchsrvr\mdbdata folder. The folder should now contain the following files:
priv.edb
pub.edb
edb.log
edb*****.log (should be a few of these)
res1.log
res2.log
The original database in renamed form (if the original database wasn’t deleted)

5 Start ONLY the Exchange Directory service.

6 Open a command prompt, and navigate to the exchsrvr\bin folder on the server.

7 Type the following command, ‘isinteg –patchE This will take time to run, and there will be no visual feedback on screen. There will however be hard disk access activity. The software is replaying all log files back into the database that were created since the database was backed up. The more log files there are, the longer this will take. Also, if the command fails, run it again straight away. It should work the second time if it failed the first time. Make sure the command completes before performing the next step. The on screen message will report, ‘The information store has been successfully updated.E

8 Start the remaining Exchange services. The priv.edb/pub.edb database will now be fully up to date up until the original problem occurred. Check this by connecting to a mailbox on the server with Outlook.