[c-nsp] Has anyone run into this error: %SYS-4-CHUNKMALLOCFAIL

[Howard Leadmon]

 Matt, that sounds like a real good possibility, as we stream lots of
video/images, to the tune of a few hundred mb/sec on a constant basis.  So I
would very much expect out packet size to be large.  

 I take it this is a fix that will have to be rolled into IOS from the sounds
of it, not something adjusting buffers will fix..  Guess I will just let it
ride as all seems to be running OK, router hasn't crashed so far (knock on
wood), and see what the next IOS update brings..

 Thanks for the reply...


---
Howard Leadmon 


> -----Original Message-----
> From: Matt Carter [mailto:matt at iseek.com.au]
> Sent: Thursday, October 25, 2007 2:47 AM
> To: 'Howard Leadmon'; 'Gregory Boehnlein'; cisco-nsp at puck.nether.net
> Subject: RE: [c-nsp] Has anyone run into this error: %SYS-4-CHUNKMALLOCFAIL
> 
> Firstly, I'm no expert in Cisco buffers, but it seems you have failures all
> the way up to and including the Huge buffer pool, which I would think is
> most certainly going to result in packet drop.
> 
> Secondly, I did come across a bug which looks awefully like the message you
> are seeing.
> 
> 006349: Jul 24 00:08:56.164 AEST: %SYS-4-CHUNKMALLOCFAIL: Could not allocate
> chunks for CEF: arp throt
> Total free: 0, Total inuse: 500, Cause : Not a dynamic chunk
>  -Process= "<interrupt level>", ipl= 1 -Traceback= 0x60635A94 0x60799A98
> 0x60799AFC 0x614551D8 0x614950A8 0x6016FC94 0x60173F28 0x6084085C 0x60773520
> 
> TAC response was i was being affected by bug CSCsj25679
> 
> Symptom: %SYS-2-CHUNKMALLOCFAIL messages on c7200 router while sending
> traffic
> 
> Conditions: %SYS-2-CHUNKMALLOCFAIL messages on c7200 router can be seen
> while box is so much stress with traffic with higher size of packets.
> (>=512bytes)
> 
> Workaround: There is no workaround. These messages not impacting any service
> like traffic and sessions drop.
> 
> Hi Matt,
> 
> The bug was first found in version 12.4(13.13)T6 and was first fixed in
> 12.4(16.13).
> 
> The developer of CSCsj25679 confirmed that your issue is the same with the
> bug.  However, he still haven't verified the mainline release on which this
> bug was first found.
> 
> <...>
> 
> By the way, developer informed me that CSCsj25679 was first seen in
> 12.4(13.10).