Hi,

I'm running DPM 2007, version 2.0.8864.0 . I have problems with many of my protection groups where the recovery points past their retention date don't delete. I did run pruneshadowcopy script but no success. Any idea what can I do besides manually deleting each one of them?

Thanks

Daniel

Silly question to ask, are the no of shadow copies beyond the specified (disk)Protection intent specified in the PG. 

 

thanks,

Parag

This posting is provided "AS IS" with no warranties, and confers no rights.

Sorry Parag but I'm not sure I understand your question. Are you asking about recovery points being older than specified? If so then yes, on many of my PG's they are older than the retention range specified. For example one of my PG's has retention range of 3 days, but my recovery points go back to 6 days ago.

Daniel 

I'm also interested in this topic, as my PG's are constantly running out of space and when you look, despite only having a protection period of 6 days, I have restore points going back 3 months+ and it will take forever to manually delete them.  Is there no way of running a script that says delete all recovery points older than x days that I can run against each PG on each server?

I found a cool powershell script that originated from MS that allows you to delete all recovery points older than x days on your DPM server.  Worked a treat!

How many snapshots per day are you taking?  One thing I didn't understand at first was that the "retention days" setting is not how many calendar days, but how many days with snapshots.  If you take snapshots say only once per week, setting your retention to "14 days", this doesn't mean you'll only have two snapshots.  You'll have 14 snapshots that extend back 14 weeks.

snapshots are daily, and on one of ourt 6 servers, it was so bad that I had to delete recovery points that were 117 days old.  Is ther eno setting I can turn on to remove these as they expire?  It's taken me 4 hours to remove everything between 20 and 117 days old for all protection groups, and now I have to do the other 5 DPM servers

My snapshots are anywhere from once to 3 times a day so this is definitely not a "calendar" issue. Jon, can you share the script you found with us?

Daniel

Ok, just checking.  Very odd -- I haven't seen this issue with DPM 2007.

Are these regular file data sources?  Or is it Exchange, SQL, or something else?

These are regular file data sources as well as C drives and SystemState, SQL, Exchange etc. I'll be honest with you, at this point I'm looking for other alternatives as DPM has been nothing but one big headache for me. And I'm not running a big network. I can only imagine what it must be for people with huge networks. My personal opinion, it's just not good enough to do the job.

D.

I completely understand. We have a large DPM infrastructure at my organization (10 DPM severs, DPM protection volumes totaling almost 200TB).  There were several challenges and it took a while to smooth out the bumps.

Hi,

 

Please try running the pruneshadowcopies.ps1 script in the DPM power shell and see if it completes without errors.  This is scheduled to run every midnight and is reponsible for deleting old recovery points beyond the retention range.  If the SQL job is not running, or if the power shell script hangs, crashes, then old RP's won't be trimmed.

Please re-open if the above suggestion do not help in resolving the issue.

Hi,

apart from bugs or failing pruning script there is an explanable answer as to why this could happen.

DPM does not purge on literal calendar age alone for very good reasons I'm sure you appreciate once understood.
Protection settings are translated to protection goals, let's take a sample: daily recovery point(s) with 3-day retention is translated to;
there should be at least 3 concise days with 1 or more valid/successful recovery points. Multiple recovery points on same day count as 1.

This means you always have the 3 most recent days with a valid recovery point regardless their literal age or distance between them.

For illustration sake; say only 1 recovery point succeeds (monday) and the other 6-days failed, before that all was fine every day.
To find 3 consecutive days with a valid recovery point we need to go back 9 days and those are kept, regardless of the 3-day retention age.
If no new recovery points are created the most recent <retention days> with at leat 1 valid recovery point are not pruned such that you always the the 'n' latest regardless of literal age.

 

 

We're getting the same thing on our Server 2012 and SCDPM 2012 SP1 installation. It's a clean installation I did last weekend. We had this issue on our Server 2008 R2 and DPM 2010 installation and after ages of troubleshooting and wasting space it turned out to be some windows updates that were preventing the prune jobs from running correctly. I checked for these updates on the new server but couldn't find them under installed updates so I'm guessing that that isn't the case. Busy running the prune script now and will see what it does.