Hello,
We have a Proxmox VE cluster (3 nodes at the moment) and are using Ceph/rbd for our primary storage. We also have some NFS shares for non-critical data, in example for ISO images and backups (our Ceph cluster is fully SSD, so not ideal for storing ISO images and backups). When Ceph is redundant, our NFS shares are not, because it's non-critical. However, when a NFS share is gone without disabling the share first in Proxmox VE, Proxmox VE becomes very slow and shows VM's black (like they are off/shutdown). In fact the VM's are still online and the VM's themself are working fine, but as admin/user working in Proxmox VE is as good as impossible because it's terribly slow and doesn't give good status information of the VM's. Even when I disable the NFS share that is down, this doesn't help at that moment. Only thing that fixes this is bringing up the NFS share again and then disable it in Proxmox VE before it's going down again (for whatever reason). If the NFS share is disabled in Proxmox VE before it goes down, there are no issues. But since this is non-redundant storage, that's not ideal.
Can this be fixed in a future release? To me this doesn't seem to be normal behavior. I can understand it will be slow when trying to access a NFS that is down (waiting for the timeout or whatever), but when I disable it in Proxmox VE it should work as before (except the NFS share is gone, offcourse) and besides that I think not whole Proxmox VE should be slow and showing VM's down when they are not.
Thank you,
We have a Proxmox VE cluster (3 nodes at the moment) and are using Ceph/rbd for our primary storage. We also have some NFS shares for non-critical data, in example for ISO images and backups (our Ceph cluster is fully SSD, so not ideal for storing ISO images and backups). When Ceph is redundant, our NFS shares are not, because it's non-critical. However, when a NFS share is gone without disabling the share first in Proxmox VE, Proxmox VE becomes very slow and shows VM's black (like they are off/shutdown). In fact the VM's are still online and the VM's themself are working fine, but as admin/user working in Proxmox VE is as good as impossible because it's terribly slow and doesn't give good status information of the VM's. Even when I disable the NFS share that is down, this doesn't help at that moment. Only thing that fixes this is bringing up the NFS share again and then disable it in Proxmox VE before it's going down again (for whatever reason). If the NFS share is disabled in Proxmox VE before it goes down, there are no issues. But since this is non-redundant storage, that's not ideal.
Can this be fixed in a future release? To me this doesn't seem to be normal behavior. I can understand it will be slow when trying to access a NFS that is down (waiting for the timeout or whatever), but when I disable it in Proxmox VE it should work as before (except the NFS share is gone, offcourse) and besides that I think not whole Proxmox VE should be slow and showing VM's down when they are not.
Thank you,
Last edited:
