Hello folks

I seem to run into the same issues all over the place.
I have a rather big backbone running on Mikrotiks. Most core networks are 100G with CCR2216.
One location is running two CCR1072 connected over a submarine cable at 310Mbit to the core with BGP4.

On BGP4 we only export a default route to that site. That site is having a wifi access network distributed over 6 sites and has OSPF routing between sites. On the core routers (the CCR1072) we every once in a while see that routing is kind of stuck. Normally there are about 100 routes in the table. On Router #2 we currently can't see any.

for example if you do

/ip route print

it just gets stuck and sits there forever.

same for

/routing/bgp/session print

Another side effect is that sometimes routes simply get missing. The router forgets about OSPF routers the other router just next to it still remembers.

The fix is simple. Rebooting the router and all will be back. However this is not a long term solution as you will have to do this again every few days.

These routers run 7.13.2 and I'm just upgrading them to 7.13.3. But I see this all the time on various routers. It seems to be that the TILE models are more affected than ARM models. To me it looks like a race condition where two cpu cores get into each others way and a deadlock occurs.

Anyone experienced similar issues in the past? It's a major stability issue for our customer base. Any things which could mitigate the issue?