Video Screencast Help

Exchange 2010 backups... Split by DB?

Created: 12 Dec 2012 • Updated: 05 Jan 2013 | 5 comments
This issue has been solved. See solution.

Hello, I'm currently facing issues with having the catch-all directive of "Microsoft Exchange Database Availability Groups:\" and the duration of backups vs log  space filling up (I have plans to resolve this - but this is going to be over Christmas break)...

If I were to split this out (still in the same policy?) to Microsoft Exchange Database Availability Groups:\DB1, Microsoft Exchange Database Availability Groups:\DB2, Microsoft Exchange Database Availability Groups:\DB3 - would this back up DB1, would it initiate a clear down of logs once once thread has completed? Just as mentioned, I have to wait for *all* of the DBs to complete before Exchange starts truncating logs...


(Note, am using GRT if that makes any difference.)

Comments 5 CommentsJump to latest comment

Marianne's picture

You may want to enable multiple datastreams and change Backup Selection to :

Microsoft Exchange Database Availability Groups:\*

This will generate a separate backup job for each db (resulting in individual truncation of logs?).

Check that Max Jobs Per Client is greater than 1.

If you want to combine the multiple datastreams with multiplexing, ensure that MPX in Storage Unit properties and in schedules are greater than 1.

Read up on Performing Exchange backups with multiple data streams in NBU for Exchange Manual.

Supporting Storage Foundation and VCS on Unix and Windows as well as NetBackup on Unix and Windows
Handy NBU Links

Richard J Cunningham's picture

Thanks Marianne, I've put this into play, hopefully it'll make a difference.. The annoying thing with issues like this is the fact it can't easily be tested, it's playing the waiting game... 7pm, will it or won't it work? :)

Also, just out of interest, what kind of speed should I be seeing from GRT backups? Just my only prior experience with GRT was using BE11 and 2k3, in which I was getting far better performance than I currently am -  backing up 800Gb + Logs is taking >24hrs and running at ~4MB\s...

Marianne's picture

No idea about your layout or individual Db sizes, but experience has shown that throughput as slow as yours that is divided into 4 streams, would generate 4 simultaneous streams at the same speed per stream, effectively cutting the backup time into a quarter of the time.

Supporting Storage Foundation and VCS on Unix and Windows as well as NetBackup on Unix and Windows
Handy NBU Links

rawilde's picture


In a given Exchange policy, if the Databases are to all be backed up on the same DAG node, log truncation is controlled by the snapshot job.  So for example, in your case if you have a single policy with DB1, DB2 and DB3 in the policy and all 3 of them are backed up on the same server, the log truncation will not be triggered until all 3 of them have been backed up and the snapshot is removed.  Even if you enable multi-streaming, that only breaks the backup jobs into multiple threads.  The snapshot process is just one thread per server.

So to accomplish what you are trying to do (log truncation at the end of each DB backup) you will need to break them into different policies.

Richard J Cunningham's picture

I've spread the DB load and enabled multi threaded backups - so hopefully this will speed our backups up and not get to the "danger zone" on the logs drive.

I'll report back tomorrow if all is well.

Cheers for your help.