Cat6509-E chaassis memory leak

Unanswered Question
Nov 12th, 2009
User Badges:

Hello :


I have six Cat6509-E switches and all of them are running IOS ver 12.2(33)SXH.

We have been observing for long time that

all the swicthes running out of memory.


And once we reboot the chassis everthing works fine.


After troubleshooting the issue , I find the following;


Nov 10 16:52:45.606: %SYS-3-CPUHOG: Task is running for (2016)msecs, more than (2000)msecs (0/0),process = Exec.(s72033_rp-ipservic


Also I see the following processes are very high preventing CPU to serve other

processes.

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

------------------ show process cpu detailed ------------------



CPU utilization for five seconds: 93%; one minute: 28%; five minutes: 12%

PID/TID 5Sec 1Min 5Min Process Prio STATE CPU

1 60.4% 11.4% 3.0% kernel 2h12m

1 6.9% 71.7% 87.7% [idle thread] 0 Ready 38d18h


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


I have found some bugs but none of them

are good for our environment.


This kind of mem leak may cause due to tarffic / drops/ errors and bearing all in my mind I have decided to capture traffic in Cat6k.


debug netdr cap rx

undeb all

sh netdr cap


I would appreciate if anyone can provide

me some way to troubleshoot it further.


BEst Regards

Arjun

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Reza Sharifi Thu, 11/12/2009 - 09:06
User Badges:
  • Super Bronze, 10000 points or more
  • Cisco Designated VIP,

    2017 LAN

Hi Arjun,


I had a very similar issue with my 6509-V-E chassis that would eventually run out of memory and reboot. The code I was running was SXI2. I upgraded to SXI2a and that seems to be ok. There are multiple bugs in the SXH and earlier SXI codes that make the switches crash.

Here are the 2 bug numbers you should take a look at


CSCsu65967

CSCsv30540


I know that these 2 bugs ware fixed in SXI3 that just was release on November 4th. So depending on your requirement and with some testing in your lab you may have to upgrade to SXI3 and see if the problem goes away.


HTH

Reza

iyde Thu, 11/12/2009 - 12:31
User Badges:
  • Silver, 250 points or more

Hi.


Notice that SXI2 has been deferred due to a catastrophic bug which causes memory leak.

SXI2a or SXI3 (the latter has passed Safe Harbor testing) is the way forward.


HTH, Ingolf

sampusarkar Fri, 11/13/2009 - 01:03
User Badges:

Hello :


Thanks for the reply but the bug you

have mentoned again it is not mayching

to our issue. I also didn't see any bug

related to CPUHOG. There are many but all

of them are affected with some commands.


That is not the case here. Do you believe

to sniff the traffic?


Regards

Arjun

Actions

This Discussion