June 1, 2006 at 3:04 am
Hi
I run scripts to restore several databases from one 2005 Server to another 2005 server. Each is a restore of a database backup followed by multiple trans logs. Two of the smaller, less active databases give me errors like "Error 18272 During restore/restart an I/O errror occurred on checkpoint file 'C:\Program files\Microsoft SQL Server\MSSQL.1\MSSQL\Backup\mdss.CKP' (operating system error 32 (error not found)). The statement is proceeding but cannot be restarted. Ensure that a valid storage location exists for the checkpoint file".
The restore seems to finish successfully after apliying further log files, the databases come online ok and DBCC doesn't report any errors.
The backup directory exists and I've granted full-control to Everyone. The same restore script always works cleanly for a bigger and more active database.
I can't find any help on MS support site - has anyone see /fixed this?
Cheers
Ian
June 1, 2006 at 3:38 am
hi,
it is possible that the transaction logs that have errors are empty ? (as they are on smaller, quieter systems). it may just be complaining about an empty log.
HTH
Paul
June 1, 2006 at 4:34 am
Paul - Good idea. All the log files are 83Kb around the window the error occur, and I have a have a task running each minute on the source server to update a heartbeat table, and the trans logs gets dumped every 5 minutes, so I don't think there are any empty log dumps.
The snippet from the error log looks like:
...
09:35:17 Log was restored... first LSN 22:459:1 last LSN 22:469:1
09:35:18 Log was restored... first LSN 22:469:1 last LSN 22:479:1
09:35:18 Log was restored... first LSN 22:479:1 last LSN 22:489:1
09:35:18 Error 18272 Severity 16 State 1
09:35:18 During restore restart an I/O error occurred...
09:35:18 Log was restored... first LSN 22:489:1 last LSN 22:499:1
09:35:18 Log was restored... first LSN 22:499:1 last LSN 22:509:1
...
Do you know if this means there's an empty log dump file after trans 22:489:1 ?
Cheers
Ian
June 1, 2006 at 4:58 am
Ian,
I don't think that there is an empty log file after 22:489:1, as it uses that for the first LSN for the next restore. I was wondering if loading an empty file was triggering a warning. I've also had a quick look around and can't find any useful info.
sorry I can't be of more help.
Paul
June 1, 2006 at 11:36 am
Paul - Think I've found a workaround - it may be related to the size of the trans log. Seems SQL Server tries to automatically checkpoint the db when the log fills past 70% (before my automated threshold sp which dumps the log when >75%). The log files for the problem dbs were only 1Mb, so I've increased them in size (5Mb) and the problem hasn't reoccurred.
Cheers
Ian
Viewing 5 posts - 1 through 4 (of 4 total)
You must be logged in to reply to this topic. Login to reply