Looping issue in Switching

[Jermy Franklin]

Hi,

 Suppose due to some loop happened in a large network, complete network got choked and administrator not able to login to core switch or other switches t identify the issue. In this scenario how to identify where is that loop happened or what is the next step we need to do?

Thanks,

Jermy

[devils_advocate]

That's a difficult question to answer, are you asking purely from a Network Troubleshooting point of view?

Can you physically access the switches to attempt to console to them?

[Jermy Franklin]

Even in console it is damn slow and switch is not responding properly?

[devils_advocate]

You need to identify where the loop is coming from and there is no 'right approach' to doing that, everyone will have a different opinion on where to begin.

A large company will (hopefully) have change control so that would be a good place to start.
Have any changes been made that day which could impact the network to potentially cause this? Its something that can be checked quickly.

Log files from your switches might be an idea, perhaps you are using SNMP to output traps to a syslog server. Any interface changes could be in the logs, may be worth a quick scan for the time the loop first occurred.

You may eventually be forced to power off kit completely and switch it on switch by switch to identify where the loop is occurring.

Its not an easy question to answer.

[Jermy Franklin]

Yeah, that's rite.

Instead of thinking where to or how to start better we can reload it. Later on by analyzing the logs of switch or we can investigate is there any new switch been introduced to the network...

Thanks for the info.,

[devils_advocate]

Reloading it will only be a temporary fix though as the loop will still likely exist so it won't take long for the CPU of the switches to start hitting the roof again.

A reload may give you enough time to pull off the config though to take a look.