LOM Interface Resets

[T_L]

Good Evening -

I am seeing a strange behavior from the LOM interfaces on a number of our gateways that I hope someone can shed some light on. We periodically lose contact with a number of our LOM interfaces (icmp/ https) and have to reset the interfaces manually via the cli.  This is not associated with any hardware issues (PDUs, power supplies and all power sensors on each box are good),  incidents, network issues, changes, etc. It is happening on series 4800, 12400, 5400, 5600 gateways - on both R77.30 and R80.20, and it seems to happen 'in bulk', or to all of them at once. Everything is good to go and then our monitoring tools alert us that the interfaces are down.  We verify that we cannot reach them, log in via ssh, start the impi services, reset mc cold (tried warm reset too) verify functionality and then stop the services - good to go again.  I upgraded the firmware for each platform per sk88064. This is not a showstopper at the moment - but has anyone else come across this?

Thanks!

[PhoneBoy]

Have you opened a TAC case by chance?

[T_L]

No sir. I honestly did not know if this was TAC worthy yet! That is my next stop.  I assume that a cpinfo on each gateway would do the trick (or at least one as a sample) as close to these events as i can get - sk92739 indicates that various system logs etc. are scooped up, although nothing specific to the LOM interfaces/cards.