[c-nsp] Strange Issue with 3560X and 4500X

[Beck, Andre]

Hi,

On Sat, Apr 12, 2014 at 01:57:42PM +0100, Antonio Soares wrote:
> It's exactly this  !
> 
> cat4k stops processing ARP, IGMP and other control protocols

Uhh-oh. Does it also stop L2 processing? In other words, is STP going
to melt down?

We've had a very similar case with 4900M boxes that slowly filled the L2
processing queue when "no vtp" was configured on an interface, on every
VTP frame they received from the peer. Great timebomb (took roughly a
fortnight to explode, and given it hit the L2 queue, it was quite the
daisycutter).
 
> CSCuj73571
> https://tools.cisco.com/bugsearch/bug/cscuj73571
> Unbelievable, this was marked with severity 2 ?!!!

What I conclude from there, this really hit public releases in 15.2(1), so
15.1(2)SG3 (aka 03.04.03.SG) isn't in danger? The dysfunctional NTP access
groups there are bad enough, but at least it seems stable otherwise...

Thanks,
Andre.
-- 
                    Cool .signatures are so 90s...

-> Andre Beck    +++ ABP-RIPE +++      IBH IT-Service GmbH, Dresden <-