High CPU usage [bonding+BGP]

Hello, our company is using CCR1036-12G-4S with 6.18 RoS.

Our configuration is:

  • on our first bonding interface have connectivity to our BGP upstream provider. This interface have two physical SFP link.
  • second bonding interface have one VLAN and two physical SFP link.

Total traffic is about 300 000pps and 1Gbps.

Our problem is with two cores, which have in peak 100% load. Usually they have about ~90%.

[admin@BGP1.dc2] /system resource cpu> print 
 # CPU                LOAD         IRQ        DISK
 0 cpu0                 0%          0%          0%
 1 cpu1                 0%          0%          0%
 2 cpu2                 0%          0%          0%
 3 cpu3                 1%          0%          0%
 4 cpu4                 0%          0%          0%
 5 cpu5                 0%          0%          0%
 6 cpu6                91%         91%          0%
 7 cpu7                 0%          0%          0%
 8 cpu8                 0%          0%          0%
 9 cpu9                 0%          0%          0%
10 cpu10               34%         34%          0%
11 cpu11               17%         17%          0%
12 cpu12                0%          0%          0%
13 cpu13                6%          6%          0%
14 cpu14                0%          0%          0%
15 cpu15                0%          0%          0%
16 cpu16                2%          0%          0%
17 cpu17                0%          0%          0%
18 cpu18                0%          0%          0%
19 cpu19                0%          0%          0%
20 cpu20                0%          0%          0%
21 cpu21                0%          0%          0%
22 cpu22                0%          0%          0%
23 cpu23                7%          7%          0%
24 cpu24                0%          0%          0%
25 cpu25                0%          0%          0%
26 cpu26                0%          0%          0%
27 cpu27                0%          0%          0%
28 cpu28                0%          0%          0%
29 cpu29                0%          0%          0%
30 cpu30               30%         30%          0%
31 cpu31                1%          0%          0%
32 cpu32                0%          0%          0%
33 cpu33               92%         92%          0%
34 cpu34                0%          0%          0%
35 cpu35                1%          0%          0%

No firewall, no Queues…

What next?

Hi,
I use full BGP table and have same problem with CCR1036-12G-4S-EM, r6.22.
mikrotik_cpu_100.png
ccr_cpu.png

Sorry to resurrect an old post but this is important to me :slight_smile:

I am looking at CCR (probably 1009 or 1012 because of redundant PSU) as an edge router:

  • 3 BGP full routing tables
  • 2 IX peers (main and backup route-servers of the IX)
  • LACP trunks
  • redistribution of default route to iBGP peers
  • simple firewall rules to prevent access to this router, connection tracking off
  • 100k pps and 300 Mbps

Are those CPU issues fixed or did you find a workaround ?
If not, do they affect packet flow ? Specifically, do they create jitter ?
What does “tool profile” show ?