Hello, everybody!
In our installation we using proxmox 4 cluster (pvecm) and shared storage based on drbd 9 and ocsf2 on top of drbd device for primary-primary mode. Cluster consist of 2 nodes.
When io load minimal, all works as we expected, i.e fine. But when users login in Windows RDS , that running on one node, io load rising up and after 3-5 min acceptable work, one of two nodes just shows kernel panic.
All that we can do in that situation is healing drbd split-brain and waiting for sync is completed.
I suppose, that this promlem is connected with FS type. May be ocfs2 is not prefered for running and synchronizing VM's? Or may be another problem presents here?
I can attach picture, that shows moment when kernel panic presents. But it is not completed log.
In our installation we using proxmox 4 cluster (pvecm) and shared storage based on drbd 9 and ocsf2 on top of drbd device for primary-primary mode. Cluster consist of 2 nodes.
When io load minimal, all works as we expected, i.e fine. But when users login in Windows RDS , that running on one node, io load rising up and after 3-5 min acceptable work, one of two nodes just shows kernel panic.
All that we can do in that situation is healing drbd split-brain and waiting for sync is completed.
I suppose, that this promlem is connected with FS type. May be ocfs2 is not prefered for running and synchronizing VM's? Or may be another problem presents here?
I can attach picture, that shows moment when kernel panic presents. But it is not completed log.