10-08-2020 10:54 AM - edited 10-08-2020 10:54 AM
We currently have Netbackup 8.2 Environment with 1 Master (W2k16 / Virtual Machine (VMWare) + 2 Media Servers (W2k16, Physicals) & Data Domain as primary disk storage target via OST. We have our Catalog backup policy configured to store via UNC Path on one of the Media Servers (Weekly Full + Daily Diffs w/a 2 Week Retention). We recently suffered a hardware failure on the physical media server that was storing the Catalog Backup. I'm thinking this is an opportunity for improvement of where to store catalog backup it for easy retrieval/recovery when needed.
One thought I had was to add a new VMDK to master server since it's a VM that way the VMDK can be easily attached to any machine and accessed in the event of a failure and not be prone to another hardware issue on the media server plus the disk can be easily grown to accommodate catalog growth.
Thoughts? What are others doing?
Thanks in advance.
Solved! Go to Solution.
10-23-2020 06:13 AM
Compared to the cost of you data, tape drives are cheap - and also protect aginst 'encryption' type malware.
10-08-2020 01:45 PM
How do you have the disaster recovery tab configured for your catalog backup?
10-08-2020 01:54 PM
Sorry if that wasn't clear in the post -- On the disaster recovery tab the path is pointed to a UNC path to one of our physical media servers
10-08-2020 01:56 PM
Do you also email it to yourself / team with attachments?
10-08-2020 01:58 PM
Yes, I receive the daily emails w/the corresponding attachments when the catalog backup has successfully completed.
10-08-2020 02:03 PM - edited 10-08-2020 02:04 PM
So you could still recover the catalog using the contents of that email.
I belive a lot of people just attach and NFS/CIFS share to their master and drop the catalog backups there. That way you could just mount it to a new server if something happens to the original server. I think giving the catalog its own VMDK is a little much but you could do it.
10-08-2020 02:08 PM - edited 10-08-2020 02:09 PM
Thanks so much for the information and reply. Right, compiling those files and emails in the event of a recovery would be possible and I always had another spot to drop the catalog backups was another recovery place to pull the files/data. Agreed, I could attach an RDM from our backend storage array and wasn't sure if that was easier/efficient than attaching a vmdk to the master server and dropping the files there as well. But NFS/CIFS another option to consider.Thanks for the feedback.
10-09-2020 01:23 PM
BasicDisk pool (attached to a drive on a server from SAN) will serve as best option to backup the Catalog for quick and easy recovery.
I would also consider having another copy on a Tape (duplicate to Tape) in case of disaster on your site, you may ship out the Tapes to offsite along with other tapes if you are already doing it.
10-14-2020 10:32 AM
10-22-2020 01:04 PM
Thanks to all for the replies.
Unfortunately, in this environment we don't have a tape unit. Any other physical servers are Application specific and don't have HBAs (not that we couldn't install etc.) except for the physical media servers but your point of planning against datastore corruption is received. I'll come up with something so I have multiple copies in addition to the daily emails received.
Thanks!
10-23-2020 06:13 AM
Compared to the cost of you data, tape drives are cheap - and also protect aginst 'encryption' type malware.