I have a Catalyst 3750 that we use as our core switch, we additionally have 3 other switches on our infrastructure.
It's been quite a while since I use the command show processes cpu, I get high % usage on core 3 of this switch.
We use Zabbix Monitoring system and we get alerts of high error rates on interface Gigabit Ethernet 1/0/22.
I have followed this guide to troubleshoot High CPU Usage:
And I was able to get these values, see attached screenshots:
FED being the process using most of the resources.
By following the guide I was able to also get to determine that said interface is causing the problem, so Zabbix is right.
What I need help on, is to continue troubleshooting this and find the root problem, as the advise of "shutting down the interface" is not an option, this interface is connected to another switch of our infrastructure.
Any other info that might help to follow on this?
follow the below correct guide :
Post below information to understand better your environment :
1. show ver
2. show switch
3. show process cpu | ex 0.00%
4. show process cpu history
Hi ! Thanks for answering!
Find attached the screenshots of what you've asked.
As far as the guide, I will follow it on Monday once I get back to the office.
I will keep this thread just in case something comes up with the screens I attached.
Hey @Oleg Volkov
Here are the screenshots that have to do with the command you asked for:
See attachments please
The switch model you have C3650 and got only 1 Switch.
I see and i believe some bugs related to your issue :
if you get a chance to try to upgrade to 3.7.5X and let us know the outcome.
Hey @Leo Laohoo back at the office!
Here is the output of "sh version" command:
What's the firmware version I have?
That will be something very difficult to do as that switch is located at the main floor, but there is another switch, actually 3 more, on another floor on the building.
All of our traffic goes to the first switch, it would be our core switch actually.
So, what are the risks of updating, downtime, and maybe other possibilities to sort it?
Are we 100% sure that this is bug-related and not something else?
I have not tried any of the possible solutions yet.
However, I was going through the Cisco article on STP and I wanted to share this screenshot with you
I ran this command on the interface having high error rates and high CPU usage:
Do you guys think this may have something to do or are we still more on-board with the fact that is a switch bug?
Logically speaking about what I read, issues with infinite STP loops will definitely make the network slower or even unresponsive.