Skip to main content

This is just a question about how we could have better way to rebuild MA. We have a windows based local ma and the data is aux copied to another media agent at a secondary site (different city);

The RAID on the local MA was returning errors; so we replaced one of the faulty disk which caused the whole raid configuration to fail.  We had to rebuild the whole virtual disk again. Unfortunately since there were only 3-6TB disks; everything OS/DDB/IndexCache/Libraries were in the same Virtual disk and we lost everything.     

 

Since we had a copy of the backup data at the secondary location; the quickest way was to rebubuild the MA.  The RAID controller was not showing any error; and it was not feasible to send another MA to the site due to lockdown   

 

So we rebilt the virutal disk again; installed/upgraded OS to Windows 2019; kept the same hostname/ip configuration; installed  Commvault agent and it made the MA online on Commserve. 

created index  cache and DDB added the new mount path in the existing library in Commvault; and sealed the old DDB. 

the backups are now running again(took a new full for all the backups; these were all file system agents) and Auxcopies running.  Is there anything else we could have done to start the backups quickly?

 

 

    

 

@Theseeker , sorry to hear about this happening, though glad you are back up and running!

One thing you could consider for next time is to promote the Secondary Copy to Primary and (temporarily) resume backups, though considering the distance, I’m not sure how much that would increase your operation window.  Once you have the main library online, run it as an Aux copy to catch up, then promote right back.

You could also consider adding in a Metallic Cloud Storage Solution (MCSS) Aux Copy (promoting to Primary) which would get you up and running VERY quickly.  It’s also a great air gapped locale to protect from Ransomware.


thanks; also I am cleaning up the old mount paths as those do not exist anymore; however it is compalning about mount path being used by DDB; although that DDB is sealed.  Am I able to delete the sealed DDB refrence? Just to tidy things up.  Just want library to point only to the new mount path

 

 


Check this thread out:

You can see which jobs are keeping it around via SQL.  Better to ensure you clear the jobs and have it age off on its own.

You CAN delete it (it’s not needed for restores, only writing and pruning), but it’s far cleaner to do it through job pruning.


Reply