DPM 2010 New Install - Creating Protection Groups - Auto, Size RRS feed

  • General discussion

  • I am starting a fresh install of DPM 2010.  Current environment is a DPM 2007 server and I have a good understanding of the size of all Protection Groups on my DPM 2007 SP1 server.

    I noticed that there were a total of 229 total file volumes as listed with diskpart (list volumes).  The Windows GUi tool Disk Manager never displays the information it always hangs.  Is this because there are so many volumes?  I have about 5 PGs with about 110 SQL databases protected, a few files sytem (tape only) and about 4 Exchange storage groups.

    Is there anyway to keep the number of volumes down?  Is there an easy way to identify filesystems that have multiple entries from the output of diskpart, list volumes?


    Friday, October 29, 2010 6:13 PM

All replies

  • Sorry the Subject line is incorrect
    Friday, October 29, 2010 6:16 PM
  • Hey Don,

    With DPM 2010 you now have the ability to place certain workload protection on 1 volume.  You can collocate workloads within one PG on one volume.

    This data collocation works for SQL, hyper-v and client protection

    As you may know, the limit for volumes in windows server is 300.  You are getting close to that limit and it is my experience that the GUI has some issues the moment you have quite a lot volumes. 

    So collocating your databases on 1 volume would be a good way to lower your volumes.

    Hope it helps


    Mike Resseler


    Visit System Center User Group Belgium @ http://scug.be and http://scug.be/blogs/scdpm
    Saturday, October 30, 2010 7:53 AM
  • But why is there a volume per item to be backed up?  I would have thought (without disk colocation) that when you create the PG, delect the sources to backup, DPM would create one volume (As seeen with the diskpart, list volumes command) per PG not 1 volume per backed up item....


    Saturday, October 30, 2010 3:27 PM
  • Hey Don,

    It actually needs 2 volumes per item.  One replica volume and one recovery point volume.

    If you want to know exactly why this is, then I suggest that you look at the movie created by Jason Buffington on how VSS actually works


    Just my 2 cents,


    Mike Resseler

    Visit System Center User Group Belgium @ http://scug.be and http://scug.be/blogs/scdpm
    Sunday, October 31, 2010 6:36 AM
  • Thanks I understand vss and the disk block changes.  What I am having a hard time with is there are two (one replica one sc) volumes per database item (instead of two per protection group.


    So when disk collocation is turned on you not only see a benefit from PG's that meet the requirements for same disk volume but you also see a benefit because all of the workloads within a single PG will share just two volumes?

    Another way to understand what I am saying is if I have a PG setup to protect SQL with 50 different server/databases that consumes about 500GB it will have 100 disk volumes.  If I delete the PG, recreate the PG using disk colocation I will probably save a bit of disk space but there will also only be 2 disk volumes allocated.

    Is that correct?  It appears that disk collocation can help reduce the number of disk volumes even with just one PG?

    Sunday, October 31, 2010 2:54 PM
  • Indeed, the biggest advantage of the collocation will be:

    1. Less volumes (and in your case with SQL, a lot less ;-)) within the PG

    2. Saving disk space because you will need less "free" space per volume.




    Visit System Center User Group Belgium @ http://scug.be and http://scug.be/blogs/scdpm
    Monday, November 1, 2010 7:24 AM