Light load, only around 1G of data traversing it
Around 3500 routes from OSPF probably only hitting a few dozen
2 OSPF neighbors
Few hundred pppoe users
No queues
10 firewall rules
7000ish connections
No L3HW offloading
2025-05-18 15:33:20 system,error,critical kernel failure in previous boot
---- current v7.18.2 2025-03-11 11:59:04 ----
2025.05.18-23:33:19.74@13: had panic
I wonder if others are running into stability issues as well? I see some posting about L3HW offloading, but we aren’t using that.
One thing I noticed is a PSU failure when the host came back up… is it possible the PSU failure caused the panic?
we had a similar problem with all our CCR1072 and 7.18.2. After upgrading, they started experiencing the following things at exactly the same time each day (around 4 in the morning) and all of them together:
dropped all pppoe connections
experienced kernel panics
sometimes rebooted
We had none of these issues when we upgraded 1036, 2116, and 2216 to 7.18.2. To resolve this we downgraded to the version that was previously installed on the 1072, 7.14.3. I don’t know if PSU problems could cause kernel panics but generally when a routeros device comes back on and one of the PSUs is disconnected it throws a log message about that PSU entering state fail. CCR2216 has 2 PSU so if one of them is disconnected during boot it will give you this log error.
Not sure if it’s the same issue, but when we were investigating the 2216 (7.18.2) as an IPoEoVlan BNG, we could reliably kernel panic it with 5 to 10 minutes of synthetic traffic: