[c-nsp] CSCsw63003 (Was: SXJ - The good, the bad, the ugly?)

Jared Mauch jared at puck.nether.net
Wed May 4 11:33:03 EDT 2011


On May 4, 2011, at 10:58 AM, Randy wrote:

> On Tue, 3 May 2011 16:41:44 +0200, Tóth András wrote
>> CSCsw63003 is indeed not fixed in SXJ yet, however it's fixed in SXI6.
> 
> Has anyone been bitten by CSCsw63003 yet?  Running SXI5 w/full tables just got
> this after 16 weeks uptime:  

We have seen it and the only software we can presently use is SXI6.

I'm not sure about your tracebacks, we would usually see a series of malloc fails and proactively reload the device.  It also helps to have long-term trending of cpu/memory usage and alerting based on some low conditions.

Having rancid login via SSH seems to consume a fair amount of memory between SSH + show run, so it's a good canary for the coal mine..

- Jared


More information about the cisco-nsp mailing list