We have a host that has played up since the removal of a DS (Not sure if this is relevant or not). It is now in a disconnected state, VMs are still online.
Should I just try restarting the management services? I need to be carefull as there is 30+ live VMs on that host.
The host has 2x1GbE onboard adapters connected for management, is this an acceptable method, or should I be using NIOC and create a virtual vmk management port group in the dvSwitch for this?
1:
You will have to restart the mangement agents.
depending on whats going on it may even take up to 2 hours to come back.
just watch it in DCUI.
you can also do a services.sh restart and watch what its doing.
also keeep pings going to some vms. ive noticed sometimes that the vms then stop responding.
then you cant do anything but hard boot.
2:
if the restart of hte management agents wont work, then you have to hard reboot the host.
I always wait till after hours in case things go bad.
Cool, I will do the restart tonight.
Now the underlying issue, how can that be tracked down? If HA was enabled we'd be in a world of hurt as the guests would evacuate off to another host, not ideal!
Appreciate your time!
HA wouldnt take the hosts down since you should have datastore heartbeats which would show the other hsots that this host is still active.