Performance degraded on 3500XL

[GStenning]

Please help.
We run a stack of 5 3500XL switches interconnected with gbic modules.
Last week one switch started to have problems, to the extent where we had to move all machines off with the exception of a few. Now there only 4 machines.
When i ping from one of the machines we get a lot of time outs.
The port lights appear to be correct. And when i plug a Laptop directly into a port the correct corresponding lights appear, but no (or hardly any) connectivity.
Can anyone please give some suggestion to try, as we have lost our network engineer (typical!)
Many Thanks

 

[tharder]

Do a "sho int count err" command. This will sho you if any of the ports are having problems You may be looking at problems that no longer exist so if there are a LOT of errors do a "clear counter" command, wait some, and check again. If the errors are incrementing, you may have a problem with cable, or duplex. We've found that many of the network problems on our switches are due to the switch/node being out of sync with the speed and duplex options. If you've had problems with this switch in the past, it may be time to replace it....

 

[vipergg]

Also look in the logs and see if the switch is complaining about anything " show log " . Do a show proc cpu and see how high this is generally running . These switches will 25-35 % even with nothing running so if it's int hat area don't worry about it . Do a sh interface and see if the ports are getting a lot of errors . Could be almost anything from an infected host to a number of other . If the switch passes POST then I would look outside the switch as the problem .

 

[lui3]

check the interfaces for errors with show interface commands or show interface counters errors

 

[GStenning]

Thanks for this.
I will give all the advice a go.
Cheers