01-21-2015 10:43 AM - edited 03-07-2019 10:19 PM
I'm troubleshooting a pair of Dell Power-Connect switches in a Dell blade chassis connected to a pair of Cisco 4900M switches. I have my 4900M switches set as spanning-tree root and backup root. The Dell switches are connected via LACP trunks to the 4900M's. Dell switch 1 to 4900 #1 and Dell switch 2 to 4900M #2. Both of the Dell switches are reporting as root switches.
I was trying to troubleshoot this yesterday and ran 'debug spanning-tree bpdu' on the primary 4900M. There was a masive amount of BPDU events scrolling by. This debug command actually took the network down. The primary 4900M was non-responsive and the secondary unit had it's CPU go to 100%. The fix was to power cycle the primary 4900M.
Why did this command take my network down?
--Patrick
01-21-2015 11:15 AM
Patrick
Were you logged in via the console ?
Jon
01-21-2015 03:57 PM
I was logged in via SSH and used 'term mon' to show the console..
--Patrick
01-22-2015 07:18 AM
Typically, the device prioritizes console output ahead of other functions. The debug spanning-tree bpdu generates a lot of output. That is what jumped the CPU to 100% and ultimately caused the device to crash.
You should be very careful with debug commands and log to the internal buffer, instead of the console.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide