Proxmox pve 9.1 slow performance

Don't use Proxmox but did you ensure the plug-in is up to date for Veeam?  I know that can cause issues.

i installed the latest pve plugin. version 13.3.0.237.

Ok was just one suggestion.  If I can think of anything else will post.

great thanks. 

Hi ​@stefan.sevasta ...have not seen any posts on here about PXMX performance degradation. Probably best to contact Veeam Support if you haven’t already. They’ll probably ask for logs for a particular job(s) with the performance issue and can hopefully provide some tips to help.

Best.

Hi, just did and already sent them some logs etc. Thanks, 

I am sure they will get you sorted out. Let us know how it goes.

as soon as i have a solution, i ll posted here.

Yes..please do so. Thank you.

same issue here.

Promox 9.1 with Ceph.

Speed ist good until veeam worker memory runs full.

increased the memory to 64gb but nothing changed. vm runs to 100% memory and poor speed.

we use currently PBS because veeam is unsubale atm

Would suggest the same as OP and contact Support to determine the issue.

Hi ​@k9sec - best bet is to contact Support at this point.

​@stefan.sevasta - Just wanted to follow up to see if you found an answer to your question. Please mark the best answer if this is resolved, including your own if needed.

Hi we opened a case in Feb and yesterday after seeing the latest logs, they said “The logs indicates that the backup process is consistently waiting on source disk read operations. This confirms the bottleneck is source read speed rather than VM-specific or worker-specific.At this stage, the evidence does not indicate a Veeam malfunction, but rather a change in source-side read performance behavior following the Proxmox 9.1 upgrade”. However from our own investigations it seems a veeam issue. While the full backup of this particular VM is being executed, we are seeing a higher IO wait on the VM itself. This is causing operational issues within the OS of the VM. Other VMs on the same host and storage do not experience high IO wait. This happens on the VMs currently being backed
up. Our storage is an external Ceph cluster. Therefore, if we had such degradation at Ceph level it would impact all VMs not just the one being backed up. Using Proxmox backup server does not have these repercussions either. Moreover we found “enableInlineMalwareScan":true” enabled. That option is disabled on our setup, but it is still enabled during proxmox backup (on vmware backup is disabled). 

Ok.  Well I would continue with support on this to get to the bottom of it.  Unfortunately no one here will be able to explain how to fix this.

That is what I am doing. Hopefully we will find a solution.