Cannot clear HA warning event

[tman24]

I was provisioning some VM's from templates in our ESX4.1 cluster last week. First VM (Win2k8 R2) deployed fine, but something very weird happened on the second one (same template).

After starting the VM, it hung about half way through the BIOS initialization screen. Absolutely nothing I could do would kill the VM, and when I finally did clear some things, I had locked resources. I google'd around, and found all the workarounds to try and free this, but eventually had to resort to migrating all the other VM's off the host and rebooting it. I tried starting the VM again, and the same thing happened (and ultimately another host reboot)!

Other than the fact I've not yet worked out why the VM hung (I'm in the process of rebuilding the templates), one other thing that happened during all this was that the cluster seemed to initiate some sort of HA event. Nothing has changed as far as I can see, and all nodes in the cluster are now working fine, but I now have a nice yellow warning triangle on the cluster icon in the VC tree view that I cannot get rid off. If I look at the error, it says;

HA initiated a failover action in cluster xxx in datacenter yyy

As far as I can tell, HA is working fine in the cluster, but I cannot seem to clear this warning. Anyone have any ideas?

[vGuy]

Hello,

have you tried restarting the management agents on the problem host:

service mgmt-vmware restart

service vmware-vpxa restart

might as well try to restart vCenter server service...let us know how it goes.

[tman24]

The ESX host itself has been rebooted twice, so it must be a VC thing. I've got some updates to do on the VC server itself, so I'll schedule in a reboot and post back.

[vGuy]

If the restart of the agents do not change the HA state, you can remove and re-add the host to the cluster. This will reinstall the HA packages and fix any misconfigurations. This worked for me last time for a similar issue.........HTH

[vPDV]

Hi

I had a similar issue, same message.

This has presumably started after having tried to destroy a template while it was being deployed. VC came up with a 'resource in use' message. After that though, the deploy task hung and had to cancel it, which did not report any issue, except from this HA message.

My current setup includes 3 vSphere 4.1 hosts in a HA/DRS cluster.

Things I tried with each host in the cluster after having gone through this post:

1. enter maintenance mode; disconnect; reconnect; exit maintenance mode

2. enter maintenance mode; remove from cluster; re-add to cluster; exit maintenance mode

I was about to restart the management agents on each host when my colleague disabled HA alltogether on the cluster, and re-enabled it. This seems to have fixed the issue.

[tman24]

Thanks, really useful info there. I don't think rebooting the actual ESX hosts is making any difference. Certainly VC is happy that HA is running in the cluster on all nodes, so totally disabling/re-enabling HA sounds like a possible solution. Nothing I've tried so far has cleared the warning message, so I've got nothing to lose!

[vPDV]

BTW I've been playing with ESX for some years now and have yet to see any management related tasks on the host that would have any negative effect on the running VM's.

[chouse · Accepted solution]

Edit the cluster configuration and uncheck "Turn on VMware HA". The alert should go away. Once the hosts have unconfigured themselves, re-enable HA on the cluster. I had this same alert and disable/re-enable HA on the cluster did the trick for me.

[tman24]

Thanks a lot. I disabled HA in the cluster, then re-enabled it. The 'error' is now cleared, and everything seems to be back to normal.

[RFSTech1]

I had the same problem and this resolved my issue also. Thanks for the tip!

[jasoncllsystems]

Try this link http://malaysiavm.com/blog/vmware-esxi-4-1-ha-warning-message/

[habibalby]

disabling and renabling HA didna't solve my problem, but restarting the managment agent on the host and restating vCenter service, cleared the Warning.

[simonebennett]

disabling/re-enabling HA also worked for me.