We're having problems monitoring UCS devices via snmpv3 (tested against 2.0(9c) and 3.0(3a)); it seems when the engine restarts (for example, if the snmp Location value is changed in the UCS software), engineTime restarts to 1 but engineBoots doesn't increment, causing notInTimeWindow faults. It's my understanding that when engineTime restarts, engineBoots should be incrememented.
What's the expected behaviour here, and what is needed for the UCS firmware to ever increment engineBoots so we can regain communication? (We've only ever recorded this value as "0" during the get-request engine discovery, and "1" subsequently. It never increments past one, which is entirely contrary to my understanding of rfc3413.)
I do have packet captures if required, however I'd prefer not to post them publicly as they're from end-user production systems.