I am trying to find a way to optimize migration times and reduce downtime of openvz containers hosted in proxmox on shared storage.
I know about the iscsi howto for openvz, which suggest a configuration with disk-over-iscsi which is completely transparent to openvz, using 1 LUN per cluster node and needing to physically copy the storage from one lun over the other via network (which is sloooowww). On my installation, where I have millions of small files the rsync approach over iscsi is not possible.
Would it be possible to use an nfs-mounted root for all openvz containers, shared accross the entire cluster? I think openvz will have some issues with quota (or perhaps not).
Any suggestions are welcome,
jinger
I know about the iscsi howto for openvz, which suggest a configuration with disk-over-iscsi which is completely transparent to openvz, using 1 LUN per cluster node and needing to physically copy the storage from one lun over the other via network (which is sloooowww). On my installation, where I have millions of small files the rsync approach over iscsi is not possible.
Would it be possible to use an nfs-mounted root for all openvz containers, shared accross the entire cluster? I think openvz will have some issues with quota (or perhaps not).
Any suggestions are welcome,
jinger