10-17-2023 12:29 PM
Hi ! (sorry for my english, google translate is my friend)
we have an instance on AIX that is about 21TB and we take it via Netbackup in an 8-channel policy. The first 8 streams are 2.5TB each and then we have other streams for archive log, control file etc... In total, we have 45 jobids for this complete backup of this instance in the NB console.
This instance has been migrated to a Linux RedHat 7.9 with the same Oracle version as the AIX.
The Linux client instance is in the same 8-channel policy as the AIX instance. It is roughly the same size, and instead of having around 45 jobids, we end up with 1,739 jobids. We don't use our own rman script; we use Netbackup's. If you have any ideas why the difference in jobs for an identical instance.
Thanks a lot!
10-17-2023 11:48 PM
Hello,
supposing you have identical policy parameters for both (especially FILESPERSET), this can be caused by Oracle storage - when AIX based Oracle DB was on non-ASM storage and Linux based is on ASM. Also supposing that you are using deduplication.
Non-ASM Oracle databases are automatically switched to so-called proxy backups to achieve better dedup ratios. These backups have also less backup jobs.
Some info here: https://www.veritas.com/content/support/en_US/doc/16226115-145903096-0/v47624000-145903096
Regards
M.
10-18-2023 02:24 PM - edited 10-18-2023 02:24 PM
Hi @laguns97
The size of the RMAN pieces is controlled by Oracle/RMAN not NetBackup so you need to look there.
From a quick google - have a check on the RMAN parameters MAXSETSIZE and MAXPIECESIZE. There may be other parameters that are alos relevant, such as the section size. This is all managed under Oracle control.
Cheers
David