SAN reboot after re-seating controller

Hello,

In the last month we twice had IMMs failing on the same controller in a 9200T PowerStore. On both accounts the active controller rebooted once the other controller got re-seated. The result was an outage of around 18 minutes and some 1200 VMs down. We were told:

“A summary of their findings indicates that there was a panic on Node A due to a race condition which can exist when there is a delay completing IO to an SSD under high load.” In short, the peer node panicked shortly after the node undergoing maintenance was re-inserted, putting the array into a state where neither Node was available to service I/O.

So at the time of the 2nd DIMM replacement we found a quiet moment but that made zero different. A full outage again.

DELL Support is stating that the issue has been mitigated in v3.6 and fully resolves in v4 but unable to provide us any official knowledgebase articles or general documentation so I obvious have doubts to whether any fixes have gone into the v3.6 release to mitigate.

Another 3000T PS had a DIMM failure however no outage occurred.

Interested to know whether other customers had experienced something similar.