Stacked Switch times out after power outage.

[efowler]

I've got two cisco switches that are stacked.  Yesterday our building lost power for 12 hours (generator also had problems, just my luck).  I got the switches both powered back up but im having problems.  When I go to ping the switch it replys but then times out a bunch.  Ive tried putty in and sometimes i can get in other times  it looses connection.  I checked show run (was able to get that to work once) and everything looked correct based on some information i had documented.  Does anyone know anything I should check to see why the stacked switch(es) keep timing out? Its like there stopping data transfer sporatically......  Thanks in advance (please help)

[Dustin Anderson]

you may need to console in, but some things you can check.

1) Assuming you have the uplink a port channel, can you disconnect one link and regain connection?

2) If you can get in, can you do a "show switch" and you should see both switches one as primary and the second as standby. If you do not see both switches or one is missing, then you may be having both switches being active and fighting over the IP causing your intermittent issue.

3) Is it only management affected, or is it disrupting all traffic? If all then I would try to power down one of the switches to see if that corrects anything.

If you can get console access, log during a reboot and look for any errors.

[efowler]

Thanks for the reply Dustin, using the show switch command I do see primary and standby.  Ive been trying to find a command to power off just one switch.  It is disrupting all traffic managment through Putty and VMware Hypervisors....

This....

Command used was "power supply on/off" - in enable mode.

SWITCH#power supply ?

  <1-9>  Switch Number

SWITCH#power supply 2 ?

  off  Set the internal PS off

  on   Set the internal PS on

Does not seem to work....

[Leo Laohoo]

Sounds like a split brain.

[efowler · Accepted solution]

Thanks all for the read/help.  I was able to figure it out.  It was a Gratuitous ARP problem.  On the Virtual Machines I had to either run a registry hack or reintiailize the network interface.  So there was no switch problem.