04-11-2022 05:53 AM
We're having a Netbackup 8.2 master server running on RHEL 6 on an old HP Proliant DL 380 G6 server. As both OS and hardware are beyond end of support, we'd like to migrate the master to a virtualized environment, because it is easier for future hardware lifecycle and upgrade.
Did someone already do this and had any issues with that ... except the statement that there maybe support issues?
04-11-2022 06:20 AM
We have done this a while ago on Windows Environment for Netbackup 7.x. I'm sure it will not be any different for 8.2
A meticulous plan is very important to complete this migration. Read all the documents from support site.
04-11-2022 11:08 AM
If you do seek more VRTS TN here these are:
04-13-2022 09:21 AM
Recovery from the catalogue procedure can be followed.
1) Keep the same hostname. Be aware that shortname and FQDN are not the same thing. And master server version must be identical.
2) Stick with the installation path and the filesystem being used. I see some acl issues with EXT3/4 -> XFS.
3) Be sure that you have full catalogue backup with master server identity (DRPGK) information. For the simplicity, backup to an STU easily transportable to the new master server. Prefer Basic Disk or AdvanceDisk.
4) Recover identity first and check to see if CA certificates are matching with the original master server.
5) Practice and document your own procedure before the actual migration.
05-20-2022 04:39 AM
Thanks. Seen that comment too late. Ran into ACL issues when restoring the catalog from ext4 to a xfs volume and the catalog recovery wizard terminated with errors.
I also noted that the BasicDisk storage units are not recreated during disaster recovery procedure. Do they need to be created manually or is it an error? The media servers, robots and medias were discovered.
05-20-2022 04:52 AM
instead recovering the database to the new server, just copy with sftp the /usr/openv of the old server to the new server. This will do the job
05-20-2022 08:07 PM
I agree with @StefanosM , as long as you add/create the web services user with the same ID, then it will work and be a simpler process than having to do a recovery.
05-22-2022 09:49 PM
You mean just install the server binaries on the new master without disaster recovery process and then rsync the whole /usr/openv from the old server over it afterwards? I think I did that 10 years ago when migrating the master server back then from Solaris to RHEL6. I wanted to do it this time the "proper" way and get rid of all the legacy stuff inside coming from previous versions back to 7.1. ;)
05-22-2022 10:24 PM
05-22-2022 10:59 PM
Some other comments, until you start running backups on the new system you can always fail back to it (you can even after starting backups but you will of course lose those).
There is no need to recreate the basic disk STU on the new master, the DR file contains the location of the catalog backup, so you just need to put the files there (if you need to use an alternate location, you can simply modify the DR file to the new correct one) - the recovery process will read the back files from the location specified (hence why it is always simpler to recover using Basic or Advanced disk).
05-22-2022 11:15 PM
Thanks. I know the document https://www.veritas.com/support/en_US/article.100022624.
The fail back I did last week already. The recovery wizard probably failed to create the basic disk storage units on the media servers because it ended with errors due to filesystem change from ext4 to xfs and many failed EXT2_IOC_SETFLAGS errors.
I'll try it from a catalog backup to disk adn ext4 formatted lvol next time to speed up recovery.