07-09-2024 08:06 AM
Hello,
We have an EMEA central WLC, Cisco 5520 in HA with firmware 8.5.171.0, hosting around 400+ APs among multiple countries and sites, I was working on the CLI for reviewing some APs logs earlier, at a moment it got stuck.
Since this moment I cannot reach the GUI anymore, and for the CLI I enter my username once, and it's stuck too.
I can only use prime and I see the WLC is still running ok, and snmp looks like to work fine.
If I go into CLI sessions in Prime, I can see my session:
Is this anything you have already faced ? Is there an easy way to solve ?
Do you know if there is a timeout for stuck CLI sessions ?
I can have access to CIMC console, but I will struggle to get into it with the JAVA old version etc ... So if you confirm it can be solved with waiting, I'm good with that !
Thanks by advance for your help.
Solved! Go to Solution.
07-10-2024 08:49 AM - edited 07-10-2024 08:49 AM
I was able to solve the issue without doing any failover.
I had a thought, I was looking into an AP log when I got stuck, and I remembered which AP is was. So I assumed the session was stuck maybe because of the AP.
So I bounced the AP port in the corresponding switch and immediately the GUI and CLI came back, Prime showing the CLI session is cleared.
I think that is a bug tbh, but I'm happy I got it cleared eventually.
07-09-2024 08:49 AM - edited 07-09-2024 09:27 AM
There is no timeout for that , well the meaning of stuck makes that contradictory. Pull the plug and check if the HA controller kicks in. If it does and the stuck controller can boot normally then you may have no interruptions. Check HA redundancy states afterwards. Also go for 8.10.196.0.
M.
07-09-2024 09:33 AM
This WLC is really critical and I would want to find another solution, we need to have this session cleared, I will try to get into the CIMC and console.
For 8.10.196.0 that's impossible as we have some old APs 3700 3600.
We are anyway planning to migrate all our 5520 to new C9800 WLCs
07-09-2024 09:48 AM
>....This WLC is really critical and I would want to find another solution,
- It's worth trying but it may not work (too) , if it's stuck at some (internal) os-level or bug,
Since you have HA , won't the standby controller take over ?
>...For 8.10.196.0 that's impossible as we have some old APs 3700 3600.
It is as per https://www.cisco.com/c/en/us/td/docs/wireless/compatibility/matrix/compatibility-matrix.html
Look them up in Table 8
M.
07-09-2024 09:59 AM - edited 07-09-2024 10:00 AM
So I can connect to CIMC after some tricks to get flash and java security bypassed, but when I connect to KVM console I have this screen :
And I cannot do anything on this console.
That is really frustrating, we just need to clear this session but we cannot access GUI neither CLI or KVM via CIMC to do it...
07-09-2024 10:13 AM
- Ref : https://community.cisco.com/t5/wireless/cisco-flexconnect-7510-won-t-boot-to-backup-image-stuck-on-quot/m-p/2706093#M95727
Not sure if that has any relevance...
M.
07-09-2024 10:43 AM
Many thanks Marce I was able to login to console using SSH and Serial over LAN feature, but even this console is stuck
07-09-2024 10:50 AM
- I then don't see any other options then pulling the plug , will the HA-standby controller take over (as it is intended to do) ?
M.
07-09-2024 10:51 AM
You mean restart the WLC from CIMC for example ? Then the standby would take over ?
07-09-2024 11:06 AM
- Yes , then on the current active controller issue the command :
show redundancy summary
M.
07-09-2024 11:31 AM
Ok and when the 1st one is back we can force the switchover to have the 1st WLC back the active.
07-09-2024 11:44 AM
- Depends , you first need to execute the command show redundancy summary
to check if there is someone (back alive) to take over (the former-stuck controller)
M.