[prev in list] [next in list] [prev in thread] [next in thread] 

List:       cisco-nsp
Subject:    [c-nsp] mld snooping defaults on 3850
From:       Sebastian Beutel <sebastian.beutel () rus ! uni-stuttgart ! de>
Date:       2016-09-30 14:51:17
Message-ID: 20160930145117.GR12813 () donquichote ! rus ! uni-stuttgart ! de
[Download RAW message or body]

Hi List,

    a couple of days ago, one of our new 3850 stacks disappeard from our
monitoring. It turned out, that it's cpu was under such heavy load that our
ping and snmp checks ran into their timeouts. Even though frame forwarding
didn't seem to be affected, even ssh to the switch was slow as several
seconds between any typed character. Going through the steps on ciscos
"Catalyst 3850 Series Switch High CPU Usage Troubleshoot" [1] tech note,
we figured out that a process with the name "NGWC L2M" was using over 90%
cpu time. Sadly i couldn't find anything more than several "tac told
us"-like statements claiming that this is correlated ti IPv6 multicasting
and that a global "ipv6 mld snooping".

So i gave it a try and achived instant relief. But some questions remain:
What actually happend? Why is this not the cisco default? Should we include
this line in our default configuration? 

Anyone ever heard of such a thing?

Best,
   Sebastian.


[1]
http://www.cisco.com/c/en/us/support/docs/switches/catalyst-3850-series-switches/117594-technote-hicpu3850-00.html
 _______________________________________________
cisco-nsp mailing list  cisco-nsp@puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/


[prev in list] [next in list] [prev in thread] [next in thread]