We are seeing several alarms being triggered for vSphere HA virtual machine failover failed on two of our clusters.
We believe that the alarms are false positives and don't want our envirnoment clutter with a bunch of false "alarms."
I have tried the following:
- Turning off HA and turning it back on (basically reconfiguring HA on the cluster)
- Disabling and re-enabling the Alarm (clearing the alarms)
Basic info on how the clusters are setup:
Cluster 1
9 ESXi 4.1 (Build 702113) hosts in vCenter 5
Cluster 2
5 ESXi 4.1 (Build 502767) hosts in vCenter 5
NOTES:
1.) Both clusters are in the same vCenter
2.) There is a known issue with latency to some of our datastores
3.) Management network is not configured with redundancy
- We don't suspect network issues on those ports
- Management ports are configured on an isolated VLAN
4.) The alarm is triggered around the time when the latency to our datastores is highest
5.) The VMs that are triggering the alarms are guests across multiple hosts in the clusters (i.e. the issue is not isolated to 1 or 2 hosts in the cluster).
Question
I understand that the datastore heartbeats may be affected by the latency, but why wouldn't the network heartbeat prevent the virtual machines from thinking an HA event has occurred?
How can we prevent these alarms from triggering? and if they are not false positives we need to figure out what is going on.
What is causing all these ugly red alarms?
