MSSQL Cluster backup - multiple instances with a single script

[ChrisJames]

Hi All, 

I am looking for a bit of help backing up a MSSQL Cluster. What we have is a 2 node active/active SQL cluster running two instances of SQL. We are looking to backup both instances using a single script referencing both instances. 

The issue I am encountering is that the backup will only work with a very limited configuration which is both instances need to be on the same node and both instances have to be added into the policy. This unfortunatly will not work because by having both clients in the policy it runs duplicate backups.

Any help?

Thanks 

Chris

[Michal_Mikulik1]

Hello,

first option is to operate NetBackup 7.7.2. Then you can use MS SQL Intelligent Policies - you can have one policy for both instances, and without need to create/maintain any scripts.

With using script method, I think your requirement is not achievable. Scripts must be local to the instance, so best saved on a shared disk and moving with the instance. So you need two.

BTW, why the requirement for one script?

Regards

Michal

[ChrisJames]

Thanks for the assistance. 

Upgrading is not currently an option but nice to know for whenever we do upgrade. 

The main reasoning for using a single script is to fit in with our existing policies and to not have to make a new policy for every instance. At the moment we have a policy which runs a script that backs up tlogs daily for example and this contains multiple servers and uses a single named script.

[Michal_Mikulik1]

Yes for multiple standalone servers, it is achievable. For cluster nodes, it is not (only for active/passive clusters where all instances would failover together).

Rgds

Michal

[ChrisJames]

Thanks for the help Michal. 

Just having a look at 7.7 and It looks be a be a pretty big overhaul on the SQL side of things. Not having to maintain scripts would be great, looks to be a step in the right direction.

Chris 

[RiaanBadenhorst]

I've had some odd results with Intellgent Policies and multiple instances. The first or default instance never works (i tested a few), but the second one works. It also spits a weird bunch of ODBC messages, even if it works :o

Aside from that, Michal is spot on with needing multiple scripts.

[Michael_G_Ander]

Only way to get only one script is to build the logic in the script, which easy can be a lot more work than just do a policy for the logical nodes in the cluster.

And you probably can't do much logic in a bch script, if any at all.

[Michael_G_Ander]

Only way to get only one script is to build the logic in the script, which easy can be a lot more work than just do a policy for the logical nodes in the cluster.

And you probably can't do much logic in a bch script, if any at all.