Upgrading backup file take a lot of time in Oracle metadata.

[antonio.amoroso]

Hi all,
After the upgrade at V12 the RMAN plungin versione v11 stopped to work.
So I proceeded to upgrade the plugin. After this I needed to upgarde metadata, stored on DataDomain.
The upgrade lasted 22 hours instead of 3/4 foreseen.
After the upgrade the duration of the job are doubled.
It's an open case with id 05991846
thanks a lot

[PetrM]

Hi Antonio,

Was it just a single run right after the upgrade when you noticed the doubled duration or it remains at this level for several days already? If it was just once after the upgrade, then I believe it's normal because more data had to be processed due to 22 hours of downtime. However, the fact that the upgrade took 22 hours looks like a technical issue, maybe related to the storage performance.

A new backup file format provides a better performance so it's not expected the backups are slower after an upgrade.

Thanks!

[antonio.amoroso]

Hi Petr,
unlucky the double duration remain at moment.
we didn't change nothing on storage configuration side, so this appear really strange
Thanks a lot
Antonio

[PetrM]

Hi Antonio,

I see that you closed the case already. Could you please tell us how did you solve the issue?

Thanks!

[antonio.amoroso]

Hi Petr,
I'm sorry but the case ID is wrong, becuase we opened two different case with similar object.
The right one is Case #06020728
Thanks
Antonio

[PetrM]

Hi Antonio,

Ok, I asked our support team leaders to prioritize your request because the behavior that you're seeing is normal at all. As far as I see our engineer waits for some additional logs, please export the requested logs to continue the investigation.

Thanks!

[wmaxime75]

I'm having the same issue. After upgrading to v12 my large rman backups more than doubled in duration. 6.5TB backup normally take 4 hours to complete. 20TB normally takes about 9 hours. Now they running in excess of 24 hours. During the backup processing rate slows to a crawl. We have to abort the backup because it's so slow. Case #06056864.

[PetrM]

Hello,

Let's wait for the results of troubleshooting from our support engineers. Looks like this problem requires our attention, we didn't see this issue in our labs during performance testing.

Thanks!

[PetrM]

Hello,

@wmaxime75 could you please clarify what type of repository do you use?

Thanks!

[wmaxime75]

2 Cisco 3260 storage servers running Win 2019. Total 7 logical drives in SOBR. Update, slowness cleared up. 1. Cleared out tons of old files left behind by rman backups. Also made a change to my NIC teams on both storage servers: Under Adv settings, disabled "Large Send Offload". Not sure that helped but speed is back to normal.

[PetrM]

Hello,

Most probably clearing out tons of old files helped to resolve it. I'll share this information with my colleagues from QA.

Thanks!