[j-nsp] MX960: High CPU FPC

[Evgeniy Aikashev]

Hi,

> Are you running any L2 services on these cards?  Could it be a layer2
> loop somewhere.
No. Only ipv4 and v6 routing. A lot of bgp sessions

> Sure, it doesn't make too much sense to be triggered by
> a RE switchover
Before a switchover, with RE0 as a primary, had no CPU problem..

> I assume the high load reappears once you've re-enabled said ports?

Yes. You are right.
For example each port of fpc2 is a member of some lags.
Aggregated interface: ae2
     xe-2/0/0                  Current   Slow periodic Collecting distributing
     xe-7/2/0                  Current   Slow periodic Collecting distributing
      xe-4/2/0                  Current   Slow periodic Collecting distributing
      xe-4/3/0                  Current   Slow periodic Collecting distributing

There are a normal CPU on all other fpcs
                     Temp  CPU Utilization (%)   Memory    Utilization (%)
Slot State            (C)  Total  Interrupt      DRAM (MB) Heap     Buffer
 4  Online            25     12          0       1024       46         31
 7  Online            24     12          0       1024       46         30

 And remains the stable when I shut  xe-2/0/0


best regards,
-- 
Evgeniy



Wednesday, February 29, 2012, 2:06:27 PM, Colin wrote:
> On 29/02/2012 9:54 PM, Evgeniy Aikashev wrote:
>> Good day,
>> I have MX960 with 2 RE.
>> Yesterday after switchover to backup RE (disk fail was the reason), I see high CPU utilization for fpc0 and fpc2  DPCE-R line cards.
>> I changed the cards, sfps but it not help. Also I found out if I disable pics xe-2/0/0 and xe-2/1/0 the CPU utilization for fpc2 decreases to normal ~10-15%.
>> Any ideas?
>>
>> evgeniy at mx960-slave>  show chassis fpc
>>                       Temp  CPU Utilization (%)   Memory    Utilization (%)
>> Slot State            (C)  Total  Interrupt      DRAM (MB) Heap     Buffer
>>    0  Online            29     86          2       1024       46         31
>>    1  Online            27     10          0       1024       46         31
>>    2  Online            25    100          3       1024       46         30
>>    3  Online            24     12          0       1024       46         31
>>    4  Online            25     12          0       1024       46         31
>>
>> Regards,
>> Evgeniy


> Are you running any L2 services on these cards?  Could it be a layer2 
> loop somewhere.  Sure, it doesn't make too much sense to be triggered by 
> a RE switchover but it sounds somewhat strange that disabling certain 
> ports fixes it.

> I assume the high load reappears once you've re-enabled said ports?

> HTH,
> Col