More strange OA3.20 / VC3.10 Firmware behaviour

[Alisdair Philp]

Following on from earlier post, we recently saw very bizarre behaviour (only affecting ESX blades)... We were doing some final proving of a new chassis to make sure HPSIM alerts came through etc and performed these tasks in a chassis with ALL firmware up to date, blades, interconnects OAs etc...

• OA switched Active to Standby (1 -->2) (F/W v3.20)
• OA switched Active to Standby (2 -->1) (F/W v3.20)
• Fibre uplink removed from Bay 2, Connection X1
• Link down confirmed
• Fibre uplink replaced into Bay 2, Connection X1
• Link up confirmed
• Flex10 in Bay 2 powered down via OA (F/W v3.10)
• Link down confirmed
• Flex10 in Bay 2 powered on via OA (F/W v3.10)
• Link up confirmed

All Windows and RHEL blades were fine - detecting down links and sending alerts happily. But... One of the 3 ESX blades we have went Profile Pending and needed to be powered down to recover as it stopped responding. The other 2 blades stayed up but their VC profiles showed no networks. When they were powered down they had to be physically removed and reinserted into their bays as they would not power on...

Strange thing is this issue from 2008 is almost the same scenarios just in a different order: http://forums13.itrc.hp.com/service/forums/questionanswer.do?admit=109447627+1288974357011+28353475&threadId=1207080

[Halil Tulumbaci]

Can you give more details about the ESX blades (ESX or ESXi Build Number and ProLiant type/generation) and the nic firmware and driver, please?