Quantcast
Channel: Data Protection Manager - File Protection forum
Viewing all articles
Browse latest Browse all 520

DPM Secondary Server Failure

$
0
0

Hello,

We've had an issue with our secondary DPM server where a planned update to SCDPM (from 2012 to 2012R2) caused some of the protected clients to not successfully create replicas. At this point I decided to restore the database which seemed like a good idea at the time as the backups were working, right up until they stopped working.

After restoring the database, none of the issues were solved and I had gained a new problem -- I could no longer edit the protection groups on the secondary server at all. 

At this point I was stuck. I couldn't get the backups to run, and I couldn't edit anything on the server to make it work. I was left with one option as far as I or Google could see -- Blow the Secondary server away and start from scratch. The secondary is only used as a DR option, so this wasn't a big deal. I removed the agent from the Primary server, uninstalled DPM from the Secondary and deleted the database. I then installed DPM2012 R2 fresh, installed the agent on the Primary from the Secondary and created a new protection group.

And herein lies the problem: I cannot see the node for "Protected Servers" listed underneath the Primary server.

Now, I know there can be all sorts of drama's related to moving protected servers around in a Primary-Secondary scenario as DPM databases tend to hold onto information. However everything I've read so far would indicate that the issue would lie on the DB of the Secondary server, which I've completely blown away and as such shouldn't be a problem. 

I have noticed that the Primary server maintains the Server ID of the Secondary server, even after it was uninstalled and removed from the environment, under the tbl_AM_Server table of it's database. I'd have thought that this should be removed upon removing the agent from the server, however it is still there. Can I manually delete this row from the database, and reinstall the agent to try and force it onto a new (and potentially correct) ID?
I've also noticed that when I install the agent onto the Primary (from the Secondary), the DPMServerID of the Primary server, as listed in the database on the Primary, does not change to point to the Secondary -- is this expected behavior?
If I do a dance from within the database and point the DPMServerID field of the two servers to each other, I can then expand the node from within the Secondary and the Primary will list the "Protected Servers" node, showing the Secondary Server. Obviously that's not what I want, as I don't want to back the Secondary up to itself, but it does show that the group can be displayed, and may be related to the Primary not believing that it has any protected servers to back up. Is there a table or a field in the DB that lists secondary servers for protection points, or any other reason why the Primary would believe that there are no valid servers for the Secondary to look after? 


Now, i thought maybe I'm missing some files or folders, so I've verified the folder and service permissions as being correct. The MTATEMP$\PSInfo folder was missing from the Primary server, so I recreated it copying the settings from the Secondary. Still no change in behavior.

I also get an enumeration error on a share when the Secondary expands to the Primary from within the protection group. Not sure if this is related? It's been happening from before the server died, so I'm thinking not.

Any suggestions on what I can look at would be greatly appreciated. It's pretty awful that we have a product that allows itself to get into this position. I know what we have done isn't exactly a "Supported Scenario", but I wouldn't have expected this much trouble.

Thanks,

Nick.


Viewing all articles
Browse latest Browse all 520

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>