VM's becomes unreachable after 4-5 hrs repeatedly.
I have kilo multi node setup with three compute nodes, VM's which are getting spawned on compute1 is all fine , but VM's which are getting spawned on compute2 and compute3 needs to soft rebooted as VM after 4-5 hours repeatedly not reachable.
What do the logs say on those compute nodes (nova-compute, syslog, dmesg)? What do the logs say inside the VMs? Is it always the same image you try to launch on different hosts or do they vary?