August 1, 2014 at 9:00 am
Hi
In the windows system error log we have started to receive the following error:
"The system failed to flush data to the transaction log. Corruption may occur" (event id 57)
I get several of these errors then:
"The default transaction resource manager on volume \\?\Volume xxxxxxxxxxxxxxx\ encountered a non-retryable error and could not start. The data contains the error code".
Then an informational level message of:
Service stopped - Source:Virtual Disk service
In the sql error log I see:
I/O Is frozen on database xxxx. No user action is required. However, if I/O is not resumed promtly, you could cancel the backup.
I/O was resumed on database xxxx. No user action is required.
Database backed up. Database:xxxx creation date , pages dumped.......
These errors have started to occur at the same time every evening on our servers which are hosted by a third party.
I am wandering what the significance of this is and how come its always at the same time ? Full database backups are not scheduled for this time yet the log says they occurred !!
August 1, 2014 at 9:39 am
The first thing I would run is this:DBCC CHECKDB (<Database Name>) WITH NO_INFOMSGS, ALL_ERRORMSGSand post the entire output. Once corruption has been ruled out then we can start investigating from there.
August 1, 2014 at 9:46 am
Keith Tate (8/1/2014)
The first thing I would run is this:
DBCC CHECKDB (<Database Name>) WITH NO_INFOMSGS, ALL_ERRORMSGS
and post the entire output. Once corruption has been ruled out then we can start investigating from there.
Just fixing the code example.
FWIW - when pasting code, a CRLF after the open code block will help prevent the conversion of < to lt and > to gt, amongst other things.
Jason...AKA CirqueDeSQLeil
_______________________________________________
I have given a name to my pain...MCM SQL Server, MVP
SQL RNNR
Posting Performance Based Questions - Gail Shaw[/url]
Learn Extended Events
August 1, 2014 at 9:55 am
August 2, 2014 at 12:25 am
Ok I have ran checkdb and the results show no errors:
DBCC CHECKDB (MyDatabase) WITH all_errormsgs, no_infomsgs executed by JoeBloggs found 0 errors and repaired 0 errors. Elapsed time: 0 hours 3 minutes 4 seconds. Internal database snapshot has split point LSN = 00050393:0006ffba:0001 and first LSN = 00050393:0006ffb9:0001.
So what is causing the i/o freeze and the subsequent database backup that the error log refers to ?
We do not have any scheduled backups at the time this happens !
I have read something about Volume Shadow Copy services which causes i/o to freeze whilst it takes a snapshot - could this be the problem ?
August 2, 2014 at 2:45 am
I/O frozen & resumed could be Shadow Copy.
Look for snapshots:
USE msdb
GO
SELECT server_name, database_name, backup_start_date
FROM backupset
WHERE is_snapshot = 1;
Maybe your hosting service is backing-up the entire VM?
If you are taking diffs, you might want to try a test restore using a full before and diff after the mysterious backup run.
Chris
August 2, 2014 at 7:52 am
looks like your third party hosts are running some sort of third party backup, have you spoken to them?
---------------------------------------------------------------------
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply