The VMware Knowledge Base provides support solutions, error messages and troubleshooting guides
VMware High Availability isolation event may attempt to start virtual machines indefinitely (1009625)
When a VMware High Availability (HA) isolation event occurs, VirtualCenter attempts to restart virtual machines on the remaining ESX hosts. As of VirtualCenter 2.5 the default behavior of restarting virtual machines due to a VMware HA isolation event was changed to leave powered on.
If the isolation response in the cluster was configured to leave powered on, VirtualCenter would attempt to restart virtual machines on other hosts even though they are still powered-on (on the host that was isolated).
As of VirtualCenter 2.5 Update 4, attempts to restart these virtual machines is limited to 5 attempts. Prior to this version, VirtualCenter would try indefinitely to restart these virtual machines.
This caused two problems:
- hostd has log spew due to VirtualCenter trying to start virtual machines, but it cannot because the disk is locked.
- Because of a potential race condition, the same virtual machines might show as registered on two or more hosts simultaneously.
The problem is resolved in VirtualCenter 2.5 Update 4.
Upgrade to VirtualCenter 2.5 Update 4 to resolve the issue.
If you want to adjust the limited attempts:
Right click on the Cluster and select Edit Settings...
Select VMware HA and click Advanced Options.
Add the following pair values to the Advanced HA options:
Option = das.maxvmrestartcount Value = <x>
where <x> 1..(reasonable number), 0 for no restarts, -1 for indefinite. The default is 5.
Disable HA at the cluster level.
Enable HA at the cluster level.
Request a Product Feature
To request a new product feature or to provide feedback on a VMware product, please visit the Request a Product Feature page.