Issue when updating B&R from 9.5.3 to 9.5.4

[mdifiglia]

Support Case #03427840
I installed B&R 9.5.3a in October of 2018 and everything went great. My backup server is a HP ProLiant DL380 Gen 9 with 4 cores, 64GB RAM and two 10GB NICS running Windows server 2016. My backups ran without issues and I never had them run over my backup window (no backups are allowed to run 7am-7pm Mon-Fri). Then I updated to 9.5.4 on 2/1/19 because we wanted to be able to replicate to Azure blob storage. Since then I have 7-12 backups failing each day. I also have jobs still running in the morning finishing merging. It seems my backup transfer rates are slower and my backup job times have doubled, tripled or more. I have called and worked with support to fix the issue. They had me create a VM proxy server which I didn't have before. This didn't help at all and I'm contemplating going back to 9.5.3. Unfortunately then I won't be able to send offsite backups to Azure Blob storage as planned. A job I have running now has the Source at 99% as my bottleneck. Proxy is 6%, Network 3% and Target 0%. Transfer rate is 50-60MB/s. Did anyone else upgrade and have this issue? I actually did a clean install of Veeam 9.5.4 to see if this would work. Unfortunately I experienced the same slowness issues..

[nitramd]

Can you add some detail, e.g., hypervisor, # of concurrent tasks on repo, number of proxies, etc.

[Gostev]

Check RAM usage at the time when backups are failing. By all indications it looks like there's not enough RAM to accommodate those added Azure offload tasks, in which case a simple fix would be to reduce the number of task slots on your backup repositories. Thanks!

[mdifiglia]

On our backend platform we are running SimpliVity Hyperconverged systems. They are running on three HP ProLiant 380 Generation 9 host servers with SSD disks (60TB total of all three) and 1TB RAM each. We use VMware 6.5 with three host servers. We have 100 VM's in our environment.

I have one B&R server that is running on an HP ProLiant with 4 core CPU, 64GB RAM. I have the SQL on another 2016 server (Full SQL install). I only have it set to run 4 concurrent tasks at this time which it is capable of running 8. I have 61 backup jobs with some jobs running backups of multiple servers. I don't have a second Proxy server right now, but I am running a VeeamOne VM server for monitoring.

Like I said in my original post everything was running great until I updated to version 4.