cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
2566
Views
0
Helpful
6
Replies

High CPU on CISCO1941

kyle woodhouse
Level 1
Level 1

Hi,

If anyone can help explain how the CPU usage on ISRG2 routers works I would greatly appreciate it. Normally when I see a CPU usage spike you can use the show proc cpu sorted command to see what process is using that CPU and then troubleshoot from there. But occasionally I have seen sustained CPU usage without any process showing its using the CPU.

Example:

CISCO1941 router

CPU spike to %99 for 1 hour

show proc cpu sorted shows no process using the cpu more than %1

My only guess is the router has ZBF configured and there is a file transfer or backup being software switched and is using all of the CPU it can get its hands on. I just don't understand why that wouldn't show in the show process cpu output as a process using the CPU.

ATLRTR-1941#show proc cpu sorted

CPU utilization for five seconds: 95%/94%; one minute: 97%; five minutes: 98%

PID Runtime(ms)     Invoked      uSecs   5Sec   1Min   5Min TTY Process

121     3545996    16326613        217  0.15%  0.07%  0.03%   0 IP Input        

  84         768         106       7245  0.15%  0.05%  0.01% 132 Virtual Exec    

  93      514168   993813653          0  0.15%  0.11%  0.08%   0 Ethernet Msec Ti

   2       86832     1586153         54  0.07%  0.04%  0.05%   0 Load Meter      

  48       63732      831382         76  0.07%  0.00%  0.00%   0 Net Background  

  50    25177852     1194371      21080  0.07%  0.03%  0.11%   0 Logger          

Message was edited by: kyle woodhouse Apparently the forums here dont support CONSOLAS font which is the best way to view router output.   Note to forum admins ADD CONSOLAS FONT SUPPORT!!

6 Replies 6

Hello

I cannot see the txt file you attached as I am posting from my iPad however You don't state what troubleshooting procedures have you done.

Have you checked the memory utilisation-what does this show.

Does this router have any LAN facing connections and have you tried temporally disabling this interface and see if the CPU utilisation drops?

Can you post the config of the router?

Res
Paul

Sent from Cisco Technical Support iPad App


Please rate and mark as an accepted solution if you have found any of the information provided useful.
This then could assist others on these forums to find a valuable answer and broadens the community’s global network.

Kind Regards
Paul

What are running on the router? BGP? OSPF? VPN? Etc, etc?

Hi, Memory utilization does not change.  I have SNMP configured and can see the graphs so i know memory isnt changing.  Will post some graphs in a second. I think disabling the internet facing interface would solve the issue but thats not the point.  The point is, why is the CPU going to 100% without a process showing me its using the CPU.  What is the point of the show process CPU command if it doesnt accuratly display processes that are using the CPU?

kyle woodhouse
Level 1
Level 1

I think my guess of it being the zone based firewall inspecting traffic is the best answer here.  I just dont understand how ZBF can use the CPU without it displaying in the "show proc cpu sorted" command.  Might open a TAC case to verify how this could possibly make sense.

Pretty clear that incoming traffic is causing the CPU spike based on the graphs.

CPU - 2 days

1.png

Packets per second - 2 days

2.png

Hi Kyle,

I dont think you provided entire output of cmd: show proc cpu sorted

Can you please uploded entire output? Also you can tune this command with "| exclude 0.00%__0.00%__0.00%"

It will exclude all lines with all zeros.

for example:

SW2#sh processes cpu sorted | exclude 0.00%__0.00%__0.00%

CPU utilization for five seconds: 6%/0%; one minute: 8%; five minutes: 8%

PID Runtime(ms)   Invoked      uSecs   5Sec   1Min   5Min TTY Process

199      892900   6503463        137  0.63%  1.16%  1.20%   0 Spanning Tree   

171        1246      5386        231  0.15%  0.02%  0.00%   0 VMATM Callback  

   3         178       132       1348  0.00%  0.13%  0.04%   0 Exec            

   4      384174     18195      21114  0.00%  1.07%  0.86%   0 Check heaps     

  34       14126       682      20712  0.00%  0.02%  0.00%   0 Per-minute Jobs 

  72         596   1211609          0  0.00%  0.01%  0.00%   0 HLFM address ret

  74        9932      4077       2436  0.00%  0.01%  0.00%   0 HULC Tcam Memory

126        3771    968057          3  0.00%  0.01%  0.00%   0 Hulc LED Process

135       13846      8166       1695  0.00%  0.04%  0.00%   0 HQM Stack Proces

231        4403      3152       1396  0.00%  0.02%  0.00%   0 IGMPSN L2MCM    

Best Regards

Please rate all helpful posts and close solved questions

Best Regards Please rate all helpful posts and close solved questions

Hi Blau,

I am very familiar with how to pipe the show proc cpu command and exclude lines that have all 0 and i did this during the event but i did not save the output.  The issue has not happened since so i cannot post output again but i assure you there were no processes that were using more than 1% of the CPU.  This is what makes no sense to me. 

I am very sure it is the zone based firewall portion of the config that is causing the high CPU because that process is stateful but i still dont understand how the cpu can be at %99 without any command to verify whats using the CPU. 

Still playing around in the lab to test this but might open a TAC case if i cant figure it out.  Thanks though!

Review Cisco Networking for a $25 gift card