cancel
Showing results for 
Search instead for 
Did you mean: 

BE DB on a SQL Cluster

nOrphf
Level 3
Hi

I have my BE 2010 DB lying on my SQL Cluster, but I pretty offent see that BE looses the connection to the DB, and my DeDupe Store goes offline and alle jobs stops.

Is it bad practice to run the DB on the Cluster, or can I change some timeout settings or something so BE doesn't loose the connection?

I'm not sure what makes BE loose the connection, as the SQL server is working fine, but it could the cluster that moves the service to another cluster node.
It gives a 10 sec timeout.

/nOrphf
5 REPLIES 5

Colin_Weaver
Moderator
Moderator
Employee Accredited Certified

Hmm do you mean you have installed BE into a Full version of SQL that is clustered  or are your talking about the way Backup Exec by default would create a "pseudo cluster" using SQL Express on both nodes (of the media server itsefl)?

If you have clustered Full SQL and it fails over then Backup Exec might not be failing over at the same time so might suffer a timeout, however if you are using the SQL Express version that it only fails over at the same time as Backup Exec.

nOrphf
Level 3
Hi

I have installed BE into a full version of SQL that is clustered

Thanks.

/nOrphf

Colin_Weaver
Moderator
Moderator
Employee Accredited Certified

In that case I would expect timeout issues (possibly more obvious if any jobs are active at the time)  if there is an interruption in SQL availability during your failover. You can have a clustered BEDB in full SQL - just be prepared for problems with active jobs and possibly the need to stop and restart the BE services as well which might be tricky if they run on a different server.

If you ever install BE as a Media Server onto a cluster and run the wizard - we don't actually cluster SQL Express - we cluster the volume where the MDF and LDF file sit (in reality we move the files) and make ssure that SQL Express on either node uses this shared location.

Then during failover the BE services stop, followed by the SQL service, then the volume is failed over and SQL is started on the other node followed by BE - at any point in time BE is never running on both nodes (apart from the remote agent) so you don't get timeouts as the services are shutdown during the failover.

nOrphf
Level 3
Is it possible yo change timeout settings i BE or would it be the best to switch to a local installed Express?

nOrphf
Level 3
Hi

I never got an answer to, wheter I should move the data base to internal database, or there is some timeout setting I could tune som the Dedupe Drive could "Survive" the move of sql service from one node to another?

/nOrphf