Today I upgraded my shiny 1072 ( v7.8 ) to 2216 ( v7.9.1 ) as my network starts to grow.
In terms of speed, nothing has been changed for now; 30 Gbps at peak times.
After migrating to 2216, I noticed high CPU even with 20 Gbps of traffic which is around 50-55% all the time while 1072 only had 5-7% CPU usage overall within the same configuration and traffic.
I have no Firewall/NAT rules at all and just have some VLANS through ports and static routes, thats all.
In terms of configuration everything is exact same with my CCR1072 but the CPU is really high I dont know why. Any advices?
But even without L3 Hardware Offloading, isn’t the CPU still very high compared to 1072? 20 Gbps traffic only consumes 5-7% CPU of 1072 while it consumes 55-57% traffic of 2216 while Fast Path is active.
in ccr1072 average cpu usage It is the result of averaging the use of the 72 cores, which tends to deliver extremely low values even when you may have some cores with a lot of load.
in ccr2216 this average calculation is done only by 16 cores which leads us in most cases to a higher more realistic value
so that numbers have a lot of room for interpretation
i think on the paper on pure CPU power ccr2216 is inferior to ccr1072
but
is easier to take advantage of ccr2216 cpu resources because they are not spread across 72 cores of only 1ghz with in-order-execution architecture
for single core tasks one ccr2216 core can do almost the same work as 3 cores of the ccr1072 because have twice the clock rate and is out-of-order execution architecture
I see, but still the CPU usage looked quite high to me compared to 1072, I checked 1072 CPUs one by one, no bottleneck even though some where at 50-60% but overall it was 12% when 2216 was at 60% on console and 71% on winbox cpu (i dont know why its also inconsistent:))
Also no idea what is “unclassified” but it also consumes reasonable amount of CPU
Thank you very much, I configured accordingly and voila CPU dropped to 0% at 17 Gbps traffic. I will be monitoring this evening when it pushes up to 50 Gbps.
Looks like VLAN interface must be set on the bridge due to Layer 2 Dependency. I was setting them directly on a port or bonding interface. After moving them to bridge and enable filtering, it all started to work. Performance is amazing now, but many drawbacks such as not able to use traffic flow etc.
I have tried this and it works ok but the challenge I am facing is related to the VLAN interfaces on the single BRIDGE. I have set the MTU of the bridge and the vlan interfaces to 9216 . However, after a device reboot, the actual MTU of the vlan interface reverts to 1500, causing my OSPF to go into an “exstart” state. To resolve this, I have to mac-telnet into the router and disable/enable the interface, which then allows the actual MTU to return to 9216.
I would like to know if there is a way to make the MTU settings on the interface VLAN persist even after a reboot, to avoid this manual intervention each time.
MTU size is inherited … bridge gets lowest MTU size of any of bridge port interfaces. Vlan interfaces get MTUs from their “trunk” interfaces (in this case from bridge). So verify what are MTU settings of all bridge ports. If they are all 9216 or higher, then you’ve encountered a bug and should contact support (preferably add supout file taken when MTU of bridge/vlan interfaces is not correct).
Thanks a lot @mkx. It will be a bug then and I have contacted support. However, I have used script/scheduler to automate bridge-restart anytime the device reboots and it works Ok for me now.
I have tried all you mentioned, and I have seen a significant drop in CPU utilization. Before the single bridge, I was pushing 2gig and CPU was around 27%, but I am seeing 12% utilization on 14gig utilization. and about 2 individual CPU cores are about 80-90%. You mentioned you are pushing 17gig and CPU is 0%. I would like that a lot. Please assist.
I do mpls/ospf/bgp/vrfs. Thanks.
