Further details are here viewtopic.php?f=3&t=149273&p=735586#p735586
But the short version is we have plenty of MikroTik routers out there all working fine with OSPF, MPLS and BGP just fine. But the ones in our data center have their SFP+ interfaces partially lock up after a few hours for no apparent reason (no logs are generated). This isn't a L3/Routing issue, it's a Layer1/OS/Network stack issue where it just totally stops sending packets, 0 packets go out that interface, only fixed by rebooting the router (for a few hours). Or turning off both OSPF and MPLS entirely
This has happened on both CCR1036 and CCR1016 platforms (tried 5 routers in total)
I've gone over the config about 2 dozen times and there's nothing really fancy in there that could cause an issue. We run CCR's elsewhere running OSPF, MPLS and iBGP. But these run eBGP with much larger BGP tables as well as being a PPPoE concentrator, the others are just routes for internal networks and summaries. And these routers are connected to a CRS317 switches with Mikrotik Direct Attach cables which we don't run that combination anywhere else in the network
I don't think the switches are a problem because they never need to be rebooted. I don't think the DAC cables are a problem if I move to another port it works, move back to original port it stops working until that router is rebooted
Has anyone else had this issue before?
System resources aren't a problem, plenty of CPU, RAM and space available. But i'm thinking maybe its just too much for the routers having so many tasks to do. Maybe we need dedicated BGP routers, dedicated PPPoE concentrators, and then another that links all together with OSPF+MPLS+iBGP