Friday, March 30, 2012

Locked files during backups on a cluster

Hello,
I am running an active-active cluster on SQL 2000 SP3a over a Windows 2003
Enterprise server. We have two production instances running (SQL1 and SQL2).
Each was setup on a different node originally. When the two instances are
running on the same node, the instance that was not setup on that node
experiences locked files and failed backups.
Here is an example:
SQL1 was setup on Node1 and SQL2 was setup on Node2. When SQL1 and SQL2 are
both running on Node1, our backup job fails because it sees that a couple of
the database backup files are locked. Even after I clear the locks and re-run
the backup script, the job still fails for the same reason. Now when I run
the same job when SQL2 is back on Node2, the job completes successfully. Not
all of the database backups fail, just the ones that are over 2Gb.
Has anyone experienced this or heard about this before? I've checked postings
here as well as TechNet, but haven't found anything.
Thanks for your assistance.
Michael
--
Message posted via SQLMonster.com
http://www.sqlmonster.com/Uwe/Forums.aspx/sql-server/200609/1> both running on Node1, our backup job fails
By 'our backup job', I assume you meant the SQL Server database backup job.
The problem should not have anything to do with how the two instances were
originally set up. Nor should it necessarily have anything to do with the
fact that the two instances were running on the same node. It was a
coincident that it did happen when the two instances were running on the same
node.
To resolve the problem, you need to first determine what process(es) or
programs are locking the backup files. My hunch is that they are most likely
locked either by an anti-virus program or a network file backup job. In other
words, the anti-virus or the network file backup job happened to collide with
the database backup on this particular node.
One way to be sure about what is locking the backup file is to add another
step immediately before the SQL Server backup to dump out all the NT handles
or just the handles on the backup file. You can use the sysinternals tool
handle.exe for this purpose.
Linchi
"michaelg via SQLMonster.com" wrote:
> Hello,
> I am running an active-active cluster on SQL 2000 SP3a over a Windows 2003
> Enterprise server. We have two production instances running (SQL1 and SQL2).
> Each was setup on a different node originally. When the two instances are
> running on the same node, the instance that was not setup on that node
> experiences locked files and failed backups.
> Here is an example:
> SQL1 was setup on Node1 and SQL2 was setup on Node2. When SQL1 and SQL2 are
> both running on Node1, our backup job fails because it sees that a couple of
> the database backup files are locked. Even after I clear the locks and re-run
> the backup script, the job still fails for the same reason. Now when I run
> the same job when SQL2 is back on Node2, the job completes successfully. Not
> all of the database backups fail, just the ones that are over 2Gb.
> Has anyone experienced this or heard about this before? I've checked postings
> here as well as TechNet, but haven't found anything.
> Thanks for your assistance.
> Michael
> --
> Message posted via SQLMonster.com
> http://www.sqlmonster.com/Uwe/Forums.aspx/sql-server/200609/1
>|||Yes, I meant the SQL Server database backup jobs.
This is not a one-time event. No matter what time I run the database backup
job on Node1, the job fails. This makes me think that it is not a tape backup
or any scheduled event. When I moved back to Node2 this morning, the job ran
fine multiple times.
I am not familiar with the handle.exe. Is this a 3rd party tool or is it
included in Microsoft?
Thanks!
Linchi Shea wrote:
>> both running on Node1, our backup job fails
>By 'our backup job', I assume you meant the SQL Server database backup job.
>The problem should not have anything to do with how the two instances were
>originally set up. Nor should it necessarily have anything to do with the
>fact that the two instances were running on the same node. It was a
>coincident that it did happen when the two instances were running on the same
>node.
>To resolve the problem, you need to first determine what process(es) or
>programs are locking the backup files. My hunch is that they are most likely
>locked either by an anti-virus program or a network file backup job. In other
>words, the anti-virus or the network file backup job happened to collide with
>the database backup on this particular node.
>One way to be sure about what is locking the backup file is to add another
>step immediately before the SQL Server backup to dump out all the NT handles
>or just the handles on the backup file. You can use the sysinternals tool
>handle.exe for this purpose.
>Linchi
>> Hello,
>[quoted text clipped - 18 lines]
>> Thanks for your assistance.
>> Michael
--
Message posted via http://www.sqlmonster.com|||This is a sysinternals tool available from www.sysinternals.com.
Linchi
"michaelg via SQLMonster.com" wrote:
> Yes, I meant the SQL Server database backup jobs.
> This is not a one-time event. No matter what time I run the database backup
> job on Node1, the job fails. This makes me think that it is not a tape backup
> or any scheduled event. When I moved back to Node2 this morning, the job ran
> fine multiple times.
> I am not familiar with the handle.exe. Is this a 3rd party tool or is it
> included in Microsoft?
> Thanks!
> Linchi Shea wrote:
> >> both running on Node1, our backup job fails
> >
> >By 'our backup job', I assume you meant the SQL Server database backup job.
> >
> >The problem should not have anything to do with how the two instances were
> >originally set up. Nor should it necessarily have anything to do with the
> >fact that the two instances were running on the same node. It was a
> >coincident that it did happen when the two instances were running on the same
> >node.
> >
> >To resolve the problem, you need to first determine what process(es) or
> >programs are locking the backup files. My hunch is that they are most likely
> >locked either by an anti-virus program or a network file backup job. In other
> >words, the anti-virus or the network file backup job happened to collide with
> >the database backup on this particular node.
> >
> >One way to be sure about what is locking the backup file is to add another
> >step immediately before the SQL Server backup to dump out all the NT handles
> >or just the handles on the backup file. You can use the sysinternals tool
> >handle.exe for this purpose.
> >
> >Linchi
> >
> >> Hello,
> >>
> >[quoted text clipped - 18 lines]
> >> Thanks for your assistance.
> >> Michael
> --
> Message posted via http://www.sqlmonster.com
>

No comments:

Post a Comment