Noticed in Virtual Center that 3 of 4 ESX hosts in a cluster were reporting alarams for health sensors. Closer insepction shows they were all getting whacky measurements for the 6  temperature sensors it is able to see from IPMI. They were reproting results like 0 degrees and -108 degrees and 115 degrees, whereas the working server is reporting 38 degrees, 40 degrees, etc.

Logged on to the System Management Homepage (i.e. the Linux/ESX HP Insight Agents are installed on these servers) and for the ones reporting broken results in VC, the temperature and fan information states were all Unknown. Looks like both Virtual Center and the HP agents are getting bad information from the hardware.

SSH’d to each ESX host and restarted SNMP and the HP agents:

service snmpd stop
service hpasm stop
service snmpd start
service hpasm start


The service startup process should be all green and if so, just reset the sensors in Virtual Centre and refresh HP SMH, everything should now be happy.

Popularity: 11% [?]