[c-nsp] Crash bug in SXH3

[Charles Spurgeon]

On Wed, Sep 03, 2008 at 02:56:13PM -0500, Church, Charles wrote:
> -----Original Message-----
> >From: cisco-nsp-bounces at puck.nether.net
> [mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of Charles Spurgeon
> >Subject: Re: [c-nsp] Crash bug in SXH3
> 
> >This bug in combination with an apparent hardware error on a Sup720
> >left one of our core routers equipped with dual sup720s crashed and
> >sitting in rommon.
> 
> Is this 'sitting in ROMMON' and not automatically rebooting normal?  We
> had a 720 do that a couple weeks ago due to a NAT bug.  I thought there
> was a crash count that put it in ROMMON only if it crashed 'x' number of
> times in a row.  Or is this one of those ROMMON bugs I saw mentioned
> recently by someone?

The TAC engineer appeared to be of the opinion that crashing to rommon
was an expected result of the hardware failure. I assumed that it was
due to crashing some number of times in a row in relation to a
hardware failure.

Our attempts to duplicate the orginal crash by repeatedly removing a
route-map were unsuccessful. We were told that the bug is intermittent
and timing related. When the TAC informed us that there also appeared
to be a hardware problem on the sup720 as well as the route-map crash
we RMA'd the sup and reverted to SXF code, so we haven't been able to
recreate the failure mode.

-Charles