I have been seeing similar things happening in my infrastructure.  We
are running mostly 4506 chassis and have seen the "unable to control
switch" message in CAM at random times.  I installed the patch a few
weeks ago, unfortunately After half of my switches lost data.  TAC has
not been of any help with this yet.  Any ideas would be welcome. 

Sean

________________________________

Sean A. Thomas, MCP, RHCT
Academic Systems Administrator
Embry-Riddle Aeronautical University
[log in to unmask]
386-226-6193 - Office
 
Any technology questions or issues, please contact IT Support at
386-226-6990

-----Original Message-----
From: Cisco Clean Access Users and Administrators
[mailto:[log in to unmask]] On Behalf Of Bill Davis
Sent: Tuesday, December 04, 2007 11:19 AM
To: [log in to unmask]
Subject: SNMP failure for OOB switches

Has anyone experienced a failure of SNMP queries to Cisco 3750 switches?

We am running Clean Access v4.1.2.0.

I posted a previous event where the Clean Access Manager was unable to
control the switches (see Nov 1, Manager loses control of OOB switches).
We have had a second occurrence of this problem last night.  The patch
mentioned in the previous thread was installed on the Managers, so we
did not lose the switch configurations in Clean Access, but the effect
on user logins was the same.

In brief, about half of our 3750 switch stacks became unmanageable and
OOB users were unable to login because the manager could not change
their vlan from auth to access.  I could see SNMP queries to/from the
Clean Access Manager, but the interface indicated the switch could not
be controlled. 
Direct SNMP queries from a work station confirmed there was no response
from the affected switches.  A reload of the switch stack did no resolve
the problem.  One of the network staff recalled a similar problem on HP
switches and the resolution was to remove all snmp-server configuration
settings on each affected switch and re-install them.  That did work for
the Cisco 3750s and we were able to bring up all the failed switch
stacks.  They have remained up so far (14+ hours).  I don't think this
is specific to Clean Access, but thought I'd post before contacting
Cisco TAC.

We are in the last week of classes and I am concerned of a repeat.  Does
anyone know of what might be causing the failure of the SNMP service on
multiple random switches? 

As a side issue:   We shut down the standby manager as a precaution when
this first occurred.  When we had brought everything back up and all
looked OK on the primary manager we booted up the standby.  When it came
up, I checked both managers using fostate.sh and they both indicated
they were primary and the other was standby.  I had to shut the second
manger back down.  The fail over for the manager has never seemed to
work.  I usually have to log into the system and restart the service to
get it to come up at all.  

Is this the standard experience for others who use a failover
configuration?

Thanks!

-Bill Davis
Housing Technology Services
Colorado State University