VMware: HA error messages appear after update of VirtualCenter 2.5 (build 64201) to VirtualCenter 2.5 Update 1 (build 84782)
Last Update: 29.05.2008
I lately saw there was an update for VirtualCenter 2.5 which was simply called
"Update 1". Since it is just an update for VirtualCenter 2.5, why
not install it? The installation/update itself went really well, it even updated
the MSSQL service without problems.
But when I then started my Virtual Infrastructure Client to connect to the VirtualCenter Server, I saw all my ESX hosts as well as the two defined Clusters with an error icon and error message. The error message says
Insufficient resources to satisfy HA failover level on ClusterXX in SG01
Unable to contact a primary HA agent in ClusterXX in SG01
I checked up on the Release Notes of this Update 1 and found the Knowledge Base Articles #1004177 "ESX Server 3i Hosts Without Swap Enabled Cannot Be Added to an HA Cluster" and #1004656 "Limited configurations are supported for VMware HA and ESX Server 3i hosts". Obviously the requirements for HA in Update 1 get more strict. But: My ESX hosts running on version 3.0.2 have swap enabled. Problem is still a fact.
When I couldn't find any solution online in the knowledge base (neither in the community forums) I decided to open a support request. VMware then gave me the following advices which are a bit odd (to manually delete installed rpm packages) but it works:
1. First of all check that all your ESX hosts can ping and resolve each other.
2. The next step is the removal of three VMware modules needed to interact with VirtualCenter (and HA). Once you reconnect the ESX hosts in a later step, the packages will be installed again automatically. We start with the removal of the vpx agent, afterwards we remove the two HA modules on the ESX machine itself (log in the service console). Do that on ALL ESX hosts having this error message.
rpm -qa | grep -i vpx
rpm -ev VMware-vpxa-x.x.x-xxxxx
rpm -qa | grep aam
rpm -ev VMware-aam-vcint-x.x.x-x
rpm -ev VMware-aam-haa-x.x.x-x
3. Now open Virtual Infrastructure Client and disconnect the ESX hosts.
4. In the Cluster settings, disable HA.
4. Connect the ESX hosts.
5. In the Cluster settings, enable HA.
This worked very fine to me!