We have recently replaced a CCR2004 (routeros 6.48.6) with a CCR1072 (6.48.6) to get more MPLS throughput. The router is part of 3 other CCR1072 routers that are connected via Layer2 links to form the backbone of the network. The router is running OSPF with MPLS (VPLS and TE).
We are consistently seeing sporadic high CPU usage on this CCR1072 with the “mpls” process being the main culprit (see profile screenshot).
The CCR2004 never gave us similar problems.
With the CCR1072 the CPU jumps to 50% usage every 30seconds and then latency increases with some packet loss.
Could this be a hardware issue?
How can we further troubleshoot?
Any inputs would be much appreciated.
hello,
you have a nice new router there 
where did you place this new ccr1072? p? pe?
i think you should contact @support for better response - don’t forget to attach some comparison screenshot between the old ccr and this new one.
good luck
Was the MPLS traffic evenly distributed among all CPU cores in the CCR2004? And is it currently being evenly distributed on the CCR1172?
According to my experience, the MPLS traffic is not distributed among the cores, and only one of the cores is in charge of processing all MPLS incoming traffic.
The router is doing PE.
We have 3x other CCR1072 routers that full similar role with same configuration and same traffic loading.
Only this one specific router in question has these CPU spikes.
Our main issue here is that we both packet loss and increased latency during the spikes. It is happening continuously throughout the day.
We are starting to suspect a hardware issue.