[c-nsp] Management interface on 2950T-24 appears to be dead

[Frank Bulk]

We monitor just CPU on that switch and it floats between 1 and 3%.  

 

We don't run SSH on that box, so I doubt it could be that.  The mgmt
interface is in our private network and not operating on VLAN 1, so it's
unlikely that it was being exploited.

 

We power-cycled the switch at 8:15 pm and by 11:20 pm it was in the same
state again.  In the meantime I did look at that switch and there was no
crashlog or anything I could see that indicated what had happened.

 

I plan to perform another reboot and upgrade it from EA11 to EA13 tonight.

 

Thanks,

 

Frank

 

From: Yann Gauteron [mailto:ygauteron at gmail.com] 
Sent: Thursday, July 16, 2009 2:42 AM
To: frnkblk at iname.com
Cc: cisco-nsp at puck.nether.net
Subject: Re: [c-nsp] Management interface on 2950T-24 appears to be dead

 

Hi Frank,

Do you monitor the switch resources (CPU, memory) ?

It is very likely that you have a memory leak in a process and no free
memory to allocate for your management process. This is at least the same
behavior that I got once with another Catalyst switch (IOS, don't remember
the version) because of a memory leak in SSH process. So when the switch
went out of free memory, it were impossible to get any administrative access
(neither SSH, nor console) to it. A reboot solved the problem. At console,
we got a message stating there were no memory free anymore. I did not play
with the mode button, so I don't know if it were working or not.

Let us know if the power cycle solved your issue, or not!

Rgs,
Y.