Hi,
We've run into performance issues with out Proxmox cluster.
Creating a linked clone has gone from being almost instantaneous to taking more than 20 seconds.
Listing the rbd's in the ceph cluster (~500 rbds) through pvesh takes even longer and trying to do the same though the API or web gui times out. This causes our provisioning service utilising the Proxmox api to fail.
Talking directly to the ceph cluster is very fast:
rbd ls -p poolname -m $hostname:6789
We're running 10 nodes. 8 of them are running 4.0-57 and we've upgrade two to 4.4-5 but the issue still exists.
Any ideas on what could be causing this?
We've run into performance issues with out Proxmox cluster.
Creating a linked clone has gone from being almost instantaneous to taking more than 20 seconds.
Listing the rbd's in the ceph cluster (~500 rbds) through pvesh takes even longer and trying to do the same though the API or web gui times out. This causes our provisioning service utilising the Proxmox api to fail.
Talking directly to the ceph cluster is very fast:
rbd ls -p poolname -m $hostname:6789
We're running 10 nodes. 8 of them are running 4.0-57 and we've upgrade two to 4.4-5 but the issue still exists.
Any ideas on what could be causing this?
