[c-nsp] Strange GEIP+ Error on 7507

[Rodney Dunn]

Wheeww...I was starting to second guess myself.

That's for the closure email.

Rodney

On Tue, Jul 25, 2006 at 05:23:36PM -0700, Carlos Sean Kamtha wrote:
> Another card with a different set of memory did the trick. Further 
> analysis revealed that 3 of the 5 cards i bought appearantly had
> bad packet memory. How odd is that? Then again they were
> were refurbished..
> 
> Thanks much for your help! :) 
> 
> On Fri, Jul 14, 2006 at 01:10:45PM -0400, Rodney Dunn wrote:
> > If you move the new GEIP+ to a new slot does the problem follow
> > it?
> > 
> > I'd try another GEIP+ I guess if it does and make sure you don't
> > swap any memory.
> > 
> > On Fri, Jul 14, 2006 at 09:48:41AM -0700, Carlos Sean Kamtha wrote:
> > > 
> > > Thought you might be interested in hearing an update.    
> > > 
> > > 
> > > The card has since crashed once. We upgraded to RSP8 last night
> > > but appears to have made no difference.
> > > 
> > > Jul 14 08:03:48.763 PDT: %VIP4-80 RM7000-3-MSG: slot4 VIP-3-PROCMEM_ECC_SINGLEBIT_ERROR: Processor memory ECC single-bit exception addr = 40038000 data = D8B00000
> > > 
> > > 
> > > In fact, they seem more frequent than before. Of course, of it take traffic
> > > completely off the router, I don't get any error messages. 
> > > 
> > > This is very strange..
> > > 
> > > On Sat, Jun 24, 2006 at 10:19:55PM -0700, Carlos Sean Kamtha wrote:
> > > > On Sun, Jun 25, 2006 at 01:11:11AM -0400, Rodney Dunn wrote:
> > > > > Ok..try it in another slot.
> > > > > 
> > > > > If you still see the probllem swap in a new RSP.
> > > > 
> > > > 
> > > > thanks! 
> > > > 
> > > > > 
> > > > > On Sat, Jun 24, 2006 at 08:57:11PM -0700, Carlos Sean Kamtha wrote:
> > > > > > On Sat, Jun 24, 2006 at 11:53:16PM -0400, Rodney Dunn wrote:
> > > > > > > Did you swap the memory between the two GEIP+'s or were they totally
> > > > > > > unique boards?
> > > > > > 
> > > > > > unique. i also reseated the memory on both cards before inserting them. 
> > > > > > 
> > > > > > > 
> > > > > > > If they were totally unique something else is causing the bit to flip in
> > > > > > > processor memory of the GEIP+ (which is VIP4-80 based which is why you see
> > > > > > > that in the log).
> > > > > > 
> > > > > > Right.
> > > > > > 
> > > > > > > 
> > > > > > > I'd say it's pretty odd that you would get a bit to flip coming from the
> > > > > > > RSP and it not be detected to/from the CyBus.
> > > > > > > 
> > > > > > > I'm hoping you reused the same memory. :)
> > > > > > 
> > > > > > i didnt. :(
> > > > > > 
> > > > > > > 
> > > > > > > On Sat, Jun 24, 2006 at 08:44:23PM -0700, Carlos Sean Kamtha wrote:
> > > > > > > > On Sat, Jun 24, 2006 at 11:37:40PM -0400, Rodney Dunn wrote:
> > > > > > > > > Please post the error again and a couple from the syslog if you have them.
> > > > > > > > > 
> > > > > > > > 
> > > > > > > > %VIP4-80 RM7000-3-MSG: slot6 VIP-3-PROCMEM_ECC_SINGLEBIT_ERROR: Processor memory ECC single-bit exception addr = 40038000 data = 00B00000
> > > > > > > >  
> > > > > > > > %VIP4-80 RM7000-3-MSG: slot6 VIP-3-PROCMEM_ECC_SINGLEBIT_ERROR: Processor memory ECC single-bit exception addr = 40038000 data = 37B00000
> > > > > > > > 
> > > > > > > > %VIP4-80 RM7000-3-MSG: slot6 VIP-3-PROCMEM_ECC_SINGLEBIT_ERROR: Processor memory ECC single-bit exception addr = 40038000 data = CCB00000
> > > > > > > > 
> > > > > > > > That's pretty much all i have. 
> > > > > > > > 
> > > > > > > > > If it says ECC it's almost NEVER a software problem.
> > > > > > > > > 
> > > > > > > > > It's a bit getting flipped somewhere and the ECC correction is detecting and
> > > > > > > > > fixing it.
> > > > > > > > 
> > > > > > > > ah.
> > > > > > > > 
> > > > > > > > > 
> > > > > > > > > If you have replaced the GEIP+ (which has ECC protection) I'd probably swap the
> > > > > > > > > RSP and see if it goes away.
> > > > > > > > 
> > > > > > > > Ok. Would I not also see the same problem with the other GEIP+ I installed
> > > > > > > > the same day?
> > > > > > > > 
> > > > > > > > > 
> > > > > > > > > Rodney
> > > > > > > > > 
> > > > > > > > > On Sat, Jun 24, 2006 at 08:28:55PM -0700, Carlos Sean Kamtha wrote:
> > > > > > > > > > On Sat, Jun 24, 2006 at 11:20:39PM -0400, Rodney Dunn wrote:
> > > > > > > > > > > Post the message again.
> > > > > > > > > > > 
> > > > > > > > > > > Please please don't change code unless you have solid proof
> > > > > > > > > > > it's a code problem.
> > > > > > > > > > > 
> > > > > > > > > > > I *thought* he said it was a single bit ECC problem that the
> > > > > > > > > > > GEIP+ was detecting. If so it could be coming from anywhere in the
> > > > > > > > > > > chassis.
> > > > > > > > > > > 
> > > > > > > > > > 
> > > > > > > > > > Are you suggesting that it could be a chassis problem? Never
> > > > > > > > > > had an issue with it until now. The chassis has been full
> > > > > > > > > > with VIP2-50s for some time..
> > > > > > > > > > 
> > > > > > > > > > Carlos.
> > > > > > > > > > 
> > > > > > > > > > > > > -----END PGP SIGNATURE-----
> > > > > > > > > > > > _______________________________________________
> > > > > > > > > > > > cisco-nsp mailing list  cisco-nsp at puck.nether.net
> > > > > > > > > > > > https://puck.nether.net/mailman/listinfo/cisco-nsp
> > > > > > > > > > > > archive at http://puck.nether.net/pipermail/cisco-nsp/
> > > > > > _______________________________________________
> > > > > > cisco-nsp mailing list  cisco-nsp at puck.nether.net
> > > > > > https://puck.nether.net/mailman/listinfo/cisco-nsp
> > > > > > archive at http://puck.nether.net/pipermail/cisco-nsp/
> > > > _______________________________________________
> > > > cisco-nsp mailing list  cisco-nsp at puck.nether.net
> > > > https://puck.nether.net/mailman/listinfo/cisco-nsp
> > > > archive at http://puck.nether.net/pipermail/cisco-nsp/