I apologize for the very basic question, but I need to understand some basic concepts about HA starting from a problem I faced some times ago.
I had a problem in a small vSphere 5 infrastructure made up by 2 ESXi 5.0 nodes.
HA is enabled and many VM are confngured to restart in case of failure of a node.
One of the 2 nodes crashed (violet screen).
While the crashed node was rebooting I noticed that the node still alive had CPU and RAM used at only 60%-70%, so there were resources to start VMs from the failed node
However, looking at the events of a typical VM on the failed node, I see the following events:
Not enough resources to failover this virtual machine. vSphere HA will retry when resources become available.
info
<date> 11:37:15
<VM_Name>
Alarm 'vSphere HA virtual machine failover failed' changed from Green to Red
info
<date> 11:37:18
<VM_Name>
Alarm 'vSphere HA virtual machine failover failed' on <VM_Name> triggered an action
info
<date> 11:37:18
<VM_Name>
Alarm 'vSphere HA virtual machine failover failed' sent email to <support email address>;
info
<date> 11:37:20
<VM_Name>
After the original nodes was rebooted, I see the following events:
Virtual machine on <host_name> is powered off
info
<date> 11:51:59
<VM_Name>
vSphere HA restarted this virtual machine
warning
<date> 11:52:33
<VM_Name>
Host is connected
info
<date> 11:54:39
<VM_Name>
Virtual machine on <host_name> is powered on
info
<date> 11:54:47
<VM_Name>
So my questions are:
- How can I see what kind of missing resource inhibited the VM from restarting on the alive node?
- Is it normal that, if a VM was unable to restart on the alive node, it is restarted on the original node when alive again?
- How can I select the VM to restart first on the alive node in there are not enough resources to restart all the VMs?
- How can I detect the overall number of VM that can be restarted on the alive node in case of the crash of a node?
- How can I set a node to automatically reboot in case of a crash (violet screen)?
Regards
marius