Below are details you are asking for:
1. We use bridge networking. Three network adapters attached to the VM use the same host bridge interface. There're no errors on this network adapter (e1000e).
2. VM was running linux-based os (customized Debian distribution).
3. We used these VMs to test HA cluster so there was Heartbeat/iSCSI/drbd (quite intensive network test). VMs involved into the cluster were set up on different physical hosts.
4. VM had been running for almost 4 days I suppose before the crash.
I looked also in libvirtd.log and found this (maybe it can help):
2013-08-13 09:31:12.654+0000: 3093: info : libvirt version: 0.9.8
2013-08-13 09:31:12.654+0000: 3093: error : qemuMonitorIORead:513 : Unable to read from monitor: Connection reset by peer
2013-08-13 09:50:59.524+0000: 3093: error : qemuMonitorIORead:513 : Unable to read from monitor: Connection reset by peer
2013-08-13 10:31:18.026+0000: 3093: error : qemuMonitorIO:603 : internal error End of file from monitor
Thanks for your help.
Below are details you are asking for: iSCSI/drbd (quite intensive network test). VMs involved into the cluster were set up on different physical hosts.
1. We use bridge networking. Three network adapters attached to the VM use the same host bridge interface. There're no errors on this network adapter (e1000e).
2. VM was running linux-based os (customized Debian distribution).
3. We used these VMs to test HA cluster so there was Heartbeat/
4. VM had been running for almost 4 days I suppose before the crash.
I looked also in libvirtd.log and found this (maybe it can help):
2013-08-13 09:31:12.654+0000: 3093: info : libvirt version: 0.9.8 ad:513 : Unable to read from monitor: Connection reset by peer ad:513 : Unable to read from monitor: Connection reset by peer
2013-08-13 09:31:12.654+0000: 3093: error : qemuMonitorIORe
2013-08-13 09:50:59.524+0000: 3093: error : qemuMonitorIORe
2013-08-13 10:31:18.026+0000: 3093: error : qemuMonitorIO:603 : internal error End of file from monitor
Regards