[c-nsp] 3750 stack(2 switches) - High CPU for an hour?

CiscoNSP List cisconsp_list at hotmail.com
Wed Dec 3 14:10:16 EST 2014

Thanks for the reply - Yes, we have also been hit with snmp/high CPU...and we have storm control on cust ports.

It coincided with a customer making some changes to there network(That connects to this switch stack) just before the high cpu issue, and resolved itself...but as I said, nothing logged (It is also configured to send snmp traps for high CPU, but no traps were sent)....anyway, 2 days of stability, then it occurred again yesterday(Only much more briefly(30seconds)....again, nothing logged, no traps...it then occurred again ~1hour later, same symptoms. I tried creating a test vlan(As it is also configured to send trap if vlan is created/deleted) - This also caused high CPU(Nothing logged, no traps, and some OSPF adjaceny up/downs)....So, did an emergency replacement of the switch(Swapped it out with a 4948).

Date: Tue, 2 Dec 2014 10:27:53 +1000
Subject: Re: [c-nsp] 3750 stack(2 switches) - High CPU for an hour?
From: adam1984 at gmail.com
To: cisconsp_list at hotmail.com
CC: cisco-nsp at puck.nether.net

We have seen high CPU Caused by SNMP.. we have implemented a view that helped. viewing the flash and license mib causes problems.

snmp-server view protect iso included
snmp-server view protect ciscoLicenseMgmtMIB excluded
snmp-server view protect ciscoFlashFileEntry excluded

snmp-server community public view protect RO

also, it might be worth implementing storm control on the client facing switchports.

On 1 December 2014 at 19:21, CiscoNSP List <cisconsp_list at hotmail.com> wrote:
Hi Everyone,

Strange situation on the weekend...started losing OSPF adjacencies to devices connected to(through) a 3750 stack (The 3750 is only doing L2, so routers connected together via the 3750)...It lasted ~1 hour (Happened while I was on a plane, so the only info I have are logs and snmp graphs)....CPU hit ~85%, traffic on all ports was "normal", PPS was "normal", switch didnt reboot and nothing at all in the logs....If the switch got hit with a mac flood or loop, I would have thought it would have a logged "something"?

Any suggestions on what else could have caused this?



cisco-nsp mailing list  cisco-nsp at puck.nether.net


archive at http://puck.nether.net/pipermail/cisco-nsp/


More information about the cisco-nsp mailing list