[c-nsp] Cisco 6500 experiencing %CPU_MONITOR-SP-6-NOT_HEARD

[j.vaningenschenau at utwente.nl]

Hi Youssef,
 
Most relevant bug ID in my archive seems to be CSCsi86691, but I'm not sure if that was "the one". The description doesn't exactly match our case.
 
We're running basic BGP with a couple of peers, but only limited routes because our SUPs don't have enough TCAM space for a full table. No MPLS or route reflector. Known problem in our environment: occasional dropping of BGP sessions where BFD is used. Can be triggered by making changes in long ACLs. We've given up on this one, our users don't notice the short drops due to redundancy. Tried TAC but we dropped the case when TAC required us to do disruptive tests.
 
This only occurs with a specific set of conditions: BGP with BFD enabled, CPU in interrupt > 10% (approx) and then modifying a standard ACL that is over 700 lines long. Determine for yourself how likely it is to hit you ;). We see it a couple of times a week, we generally lose one BGP session for 5 - 15 seconds.
 
Perhaps others who use BGP+RR+MPLS know more important caveats...
 

Regards,
 
Jeroen van Ingen
ICT Service Centre
University of Twente, P.O.Box 217, 7500 AE Enschede, The Netherlands


________________________________

From: Youssef Bengelloun-Zahr [mailto:youssef at 720.fr] 
Sent: woensdag 16 juni 2010 15:10
To: Ingen Schenau, J. van (ICTS)
Cc: cisco-nsp at puck.nether.net
Subject: Re: [c-nsp] Cisco 6500 experiencing %CPU_MONITOR-SP-6-NOT_HEARD


Hello Jeroen,

Thanks for the feedback. If you can find the bug IDs, please do not hesitate to send them, it come in handy sometimes.

I have been thinking of upgrading to SXI3 (why not SXI4, hey ;-) for a long time and have labed it, all my configs were correctly accepted.

We are running some basic BGP / MPLS and route reflection on this router, have experienced any weird things regarding theese on SXI3 / SXI4 ?

Thanks again.

Best regards.

Y.




2010/6/16 <j.vaningenschenau at utwente.nl>


	Hi Yousef,
	

	> Just for the record, I will post this in case some guys out there
	> have the
	> same problem some day.
	>
	> Last friday, one of my core routers, a Cisco 6509 with two SUP720-3BXL
	> modules running s72033-advipservicesk9_wan-mz.122-33.SXH2a, crashed
	> and
	> restarted out of the blue.
	>
	> Crashfile info says the following :
	
	
	-=snip=-
	

	> Personally, I'd say I hit a bug with this but I can't seem to find it
	> using
	> cisco web tools. Anyone could point me to the right direction ?
	
	
	We had similar crashes in 2007-2009 (on SUP720-3B). After several *long*
	TAC cases, it turned out that we hit a couple of bugs. I can't find the
	bug IDs at the moment, but according to my email archive, the fixes were
	included in SXH4.
	
	I'd recommend trying a more recent SXH build as a lot of issues have
	been fixed since SXH2a. Or, if you're comfortable with bigger upgrade
	steps: we're running SXI3a now, which has been more stable in our
	environment than SXH has been. However, as always with bugs & features
	in IOS, YMMV.
	
	By the way, the root cause in our case had to do with interrupt masking;
	it was mainly triggered by non IP packets from directly connected
	network segments. Our case was only reproducible by replaying actual
	traffic captures, not with "synthetic" IP traffic.
	
	
	Regards,
	
	Jeroen van Ingen
	ICT Service Centre
	University of Twente, P.O.Box 217, 7500 AE Enschede, The Netherlands
	




-- 
Youssef BENGELLOUN-ZAHR ......................................................
Ingénieur Réseaux et Télécoms


Technopole de l'Aube  en Champagne - BP 601 - 10901 TROYES  Cedex 9
Agence Paris : 6, rue Charles Floquet - 92120 MONTROUGE
Tel                 +33 (0) 825 000 720
Tel. direct      +33 (0) 1 77 35 59 14
Tel. portable  +33 (0) 6 22 42 63 80
Email            ybz at 720.fr
...............................................................................................www.720.fr