HIGH CPU with VSH Process in Module ACE 30

[rc5295509]

Hi

In a ace modulo, the CPU is 60 % aprox. and I a process "VSH" that generate this high use of CPU, some session lost.

How can control this process ?

version:

Version A5(1.2)

switch/prod# sh processes cpu | exclude 0.0

CPU utilization for five seconds: 100%; one minute: 59%; five minutes: 58%

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

-----  -----------  --------  -----  -----   ----    ----    ----  -----------

10810      79264593  164468057    481     47  71.18%  19.54%  20.34%  vsh

22988      38268195  92738856    412     57   0.1 %  22.51%  19.42%  vsh

23273      38439564  91513671    420      6  71.20%  24.51%  22.73%  vsh

24133      38732780  89694759    431     37   0.5 %  15.25%  16.59%  vsh

switch/prod#

if you know somenthing..  please,  let me know

[Kanwaljeet Singh]

Hi,

It could be due to stuck VSH sessions. Are you using any scripts which login to ACE periodically for collecting some outputs?

Do you see it going above 60%. At that %age you should not see any session loss. Are you seeing any denied counters in sh resoource-usage?

This is a complicated process and TAC might need to go to debugshell for troubleshooting the problem. I would suggesting collecting show tech and syslogs during the high CPU usage and open a TAC case for further investigation.

It could be related to below DDTS: CSCua90111:   ACE not clearing VSH process resulting in high CPU usage.

Regards,

Kanwal

[rc5295509]

Hi

there are not script..

but the ouput command "sh resource usage" have denied

to-cp-ipcp rate            1624       4321          0       5000     306795

I´d open a case in TAC.   you know what is this resource(to-cp-ipcp)?

[Kanwaljeet Singh]

Hi,

to-cp-ipcp-Limits represents the IPCP traffic from the DP(dataplane) to the CP(Control Plane) in
packets per second. This keyword prevents the overwhelming of the CP under high syslog
rate conditions (for example, level 7 messages).

This is an umbrella resource intended to prevent CP saturation so this should not impact
the traffic on ACE. One way to mitigate is to reduce the logging level on each context to
limit syslog messages coming from DP to CP.
 
 So if you have a high logging level configured in contexts like logging level 7(debug) please reduce it 
 to lower level like "Warning". High level logging is recommended when you are troubleshooting an issue and then
 you can turn it to lower level again.
 
Also, internal research threw up that DDTS but i would still suggest opening a TAC case for thorough investigation.
 
 
 
 Regards,
 Kanwla

[rc5295509]

hi kanwal

The bug make match with the symptom..

Thanks

Rodrigo