Switches become unavailable for management
Hi.
There are SG3452XP switches (about 20 devices), adopted by Omada SDN (version 5.15.24.19).
Sometimes one can become unavailable for management - status in Omada becomes "disconnected" and it's impossible to login.
Can happen with any switch, and resolved by reboot.
While unavailable, they still continue working, they even answer ping. If you try to connect by console, they ask login/password, but hang up after you enter them.
We noticed that before problems start the utilization of CPU boosts from usual 6% up to 53%. They work for about 14 hours with that 54% utilization and eventually become unavailable.
Have you experienced same problems? What can be the solution?
- Copy Link
- Subscribe
- Bookmark
- Report Inappropriate Content
Thank you for your post. When the switch disconnects, please observe the port LEDs and the SYS LED—are they solid or flashing? Kindly collect the device info for analysis. If the SYS LED stays solid, the switch may have crashed; after rebooting and restoring service, run the CLI command show dump info and provide the output for further investigation.
- Copy Link
- Report Inappropriate Content
So, in our case the solution was to disable 802.1X and built-in RADIUS-server on Omada SDN.
After being done, there were no switch hang-ups.
Hope this helps someone!
- Copy Link
- Report Inappropriate Content
Thank you for your post. When the switch disconnects, please observe the port LEDs and the SYS LED—are they solid or flashing? Kindly collect the device info for analysis. If the SYS LED stays solid, the switch may have crashed; after rebooting and restoring service, run the CLI command show dump info and provide the output for further investigation.
- Copy Link
- Report Inappropriate Content
@Ethan-TP Thank you for answering.
All LEDs are flashing.
I tried to run "show dump info", but there is no such command.
- Copy Link
- Report Inappropriate Content
We are facing the same problem as mentioned in this thread:
- Copy Link
- Report Inappropriate Content
When the issue re-appears, click on the device and, in the right-hand pop-up panel, locate “Device Info”; please export it and send it to us for analysis.
In addition, while the problem is occurring, access the device by SSH or console and collect the output of the following commands:
show running-config
show logging buffer
show cpu-utilization
show memory-utilization
show system-time
show system-time ntp
debug show ecs log
debug show ecs controller-conf
debug show ecs set-cfg 1
debug show ecs set-cfg 2
debug show ecs set-cfg 3
debug show ecs set-cfg 4
debug show ecs set-cfg 5
You may also try pinging the Controller IP address.
- Copy Link
- Report Inappropriate Content
When the issue occurs, the switch becomes unavailable. - I wrote before.
So, it's impossible to manage via Omada SDN, no opportunity to export Device-Info. Also, no opportunity to access the device by SSH or Console.
- Copy Link
- Report Inappropriate Content
So, in our case the solution was to disable 802.1X and built-in RADIUS-server on Omada SDN.
After being done, there were no switch hang-ups.
Hope this helps someone!
- Copy Link
- Report Inappropriate Content
Information
Helpful: 0
Views: 172
Replies: 6
Voters 0
No one has voted for it yet.