Hello, this is weird.
we have a CCR1072 still in production as a border router in a regional IX. It runs latest 6.49 (although this happened with different versions), does one BGP full view plus about 50 smaller peers, and pretty much nothing else except for IGP and input (-> to the router) firewall.
It happened to us three times now that, out of a sudden after months of working fine, it would stop sending traffic from *any* interface except the SFP+ 1 (which in this case is the peering lan of the IX).
According to our post-event data collection things go like this:
- the router, out of nothing, reboots itself, without anything in the logs. Connected devices just see all connecting interfaces go down, stay down enough for a reboot, then up again.
- after this unexpected reboot, not a single ethernet frame is outputted from the router by any SFP interface except sfppp1, there, the router keeps all peering sessions up and appears completely normal. All interfaces are mixed copper 10GB/s DACs and actual SFPs, doesn't seem to be a factor.
- no amount of reboot, power cycle, disconnecting cables, changing cables, SFPs, rebooting adjacent routers, can get it working again. Sometimes, after a full power cycle (not a soft reboot) it appears to be working for 20-30s then locks up all the interfaces again.
- after a few hours, and not before, a final power cycle clears it.
After the second episode, we replaced it with a similar CCR1072 keeping the same config. Worked fine for months, the new one did the exact same just yesterday.
Note that all this only happened in production, where we don't have any control on the traffic going through, so it *could* be some kind of peculiar traffic causing the lock up, that disappears in a few hours.
Here: viewtopic.php?t=102946 there's a mention of something similar.