Long story short:
We had 3 proxmox servers in a customer environment all running v1.9
We had an issue with backups to a NFS filer taking a long time that we posted both here and purchased a support ticket for. The 'fix' was to move to the following kernel:
atom1:/# pveversion -v
pve-manager: 1.9-26 (pve-manager/1.9/6567)
running kernel: 2.6.35-2-pveThis did fix the problem... However we upgraded one host to proxmox 2.1 and now we have the issue again. What makes it worse is that once we do a backup from the v2.1 server that starts to take a long time, it seems to actually affect the openfiler server itself as the load average goes high and we can only get it back by removing all NFS connections to regardless of whether there is traffic or not. ie. we remove the storage from the config on all the proxmox servers and then put them back and all seems well.
So whatever the issue is in Kernel 2-6.32 vs 2-6.35... we have the issue again.
I'm hoping there is a fix to this as it stands right now we need to actually downgrade the server back to v1.9 and ultimately we are looking for some of the new functionality in v2.1
As a side note when we went to upgrade the box originally to v2.1 from 1.9 it failed miserably and I had to actually install 2.1 on the server and restore the VMs from backup.
Also when I say a long time... I mean a backup that normally takes 1.5 hours in v1.9 took over 24 hours to complete.
We had 3 proxmox servers in a customer environment all running v1.9
We had an issue with backups to a NFS filer taking a long time that we posted both here and purchased a support ticket for. The 'fix' was to move to the following kernel:
atom1:/# pveversion -v
pve-manager: 1.9-26 (pve-manager/1.9/6567)
running kernel: 2.6.35-2-pveThis did fix the problem... However we upgraded one host to proxmox 2.1 and now we have the issue again. What makes it worse is that once we do a backup from the v2.1 server that starts to take a long time, it seems to actually affect the openfiler server itself as the load average goes high and we can only get it back by removing all NFS connections to regardless of whether there is traffic or not. ie. we remove the storage from the config on all the proxmox servers and then put them back and all seems well.
So whatever the issue is in Kernel 2-6.32 vs 2-6.35... we have the issue again.
I'm hoping there is a fix to this as it stands right now we need to actually downgrade the server back to v1.9 and ultimately we are looking for some of the new functionality in v2.1
As a side note when we went to upgrade the box originally to v2.1 from 1.9 it failed miserably and I had to actually install 2.1 on the server and restore the VMs from backup.
Also when I say a long time... I mean a backup that normally takes 1.5 hours in v1.9 took over 24 hours to complete.