[PVE-User] Nodes lose connection to master

[Steve Weber]

Hello,

we have a cluster of three nodes:
                mox-110 (slave)
                mox-114 (master)
                mox-118 (slave)
Yesterday morning all three nodes shown in syslog: "crit: cpg_send_message failed: 9".
After opening the WebGUI of one node after the other, only the active node was shown
as online the others were shown as offline.

At evening I restarted the master and the error disappeared in his log.
At the slaves the syslog entry "crit: cpg_send_message failed: 9" was visible again.
But the master recognized changes at the pve-worker processes.

è I restarted the slaves and everything went back to normal.

The problem resulted in my opinion by a malfunction of the pve-cluster service.

Is there any way to prevent this from happening again?

Greetings
Steve

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.proxmox.com/pipermail/pve-user/attachments/20130206/de36832d/attachment.htm>