Performance issues after upgrading to 9.5 U4a

[stewsie]

Since upgrading last week we have so many jobs that are now taking a long time to merge or create full backup maps for tape jobs. After upgrading last Monday, jobs started to slow up over a number of days. Some merge operations are now taking days to complete and then the process starts again when the job next runs. Storage has been checked and confirmed there are no issues. We are using CIFS and I know this isn't the best option but this has been configured for almost 5 years and this is the worst performance we have experienced

I have a support call open (03525121) and was told this appears to be a known issue and I was advised to replace a .dll file which I did. There has been no real improvement.

Anyone else having issues like this? Also have an issue trying to restore an SQL database which is currently sat at more than 40 hours and appears to be doing nothing. May not be related to the upgrade and I have another support call open for that

[Dima P.]

Hello Paul,

Thank you for your post. We are aware of the mentioned issue and RnD folks are still investigating why the proposed fix had no effect. Kindly keep working with our support team and let us know how the investigation goes from your end. Cheers!

[stewsie]

Working with the Veeam engineer I have assigned proxies to backup repositories and the performance overnight has been the best since the upgrade. I am still monitoring the service and will update the engineer but at the moment performance is much better.

I had assigned proxies to repositories before (maybe a couple of years ago) but it never seemed to make any difference and they were changed back to auto. This time however I have noticed a change. I always though the backup server was overloaded during the merge operation.

https://eur02.safelinks.protection.outl ... reserved=0

[stewsie]

The performance issue now appears to be resolved following the work I carried out assigning proxies to repositories rather than leaving as auto which had been the case for almost 5 years.

This was recommended by Evgenii Ivanov who has been incredibly helpful in resolving this issue. The support call is still open as I understand investigations are still ongoing following the upgrade but performance is now back to the levels prior to the upgrade.