I have 2 WLC 5760 stacked. Everything was working fine. We faced some issues regarding authentication for an SSID that use radius authentication. We noticed on (show radius statistics) that we have a high timeout number of request. So we changed the default radius-server timeout from 5 to 20 sec. After that with some times we noticed that when we use show run, no output never shows and show tech shows only few lines and becomes unresponsive, other shows all fine and ok. At this instance we could show start-up config. Then we reverted back the radius timeout to 10 sec and while saving the configuration, it never shows it saved and become unresponsive too. Then we couldn't make show start-up as previously and it showed this message
Cisco-WLC-5760#show startup-config Using 0 out of 2097152 bytes, uncompressed size = 7809 bytes %Error opening nvram:/startup-config (Device or resource busy)
After that we noticed the below log message using terminal monitor on telnet:
%PARSER-6-WMLRETRY: Write memory lock currently held by pid '404', automatic retry. -Process= "Virtual Exec", ipl= 0, pid= 406
Then after a while, we couldn't initiate any telnet sessions to the WLC however it is pinging normally (The remote system refused the connection). The WLC runs IOS version 3.3.4SE
Notice, all wifi services working fine however this issue appeared. Any suggestions?
I had a similar behaviour with Cat4500E. TAC told me to upgrade the IOS because it looked like a bug. Usually, a reboot would help as an immediate remedy. We checked the memory, the processes, the lines and all looked ok.
I have rebooted and it was ok and after that i upgraded to 3.3.5SE. It looks fine now but we never knew what is the trigger of this bug to happen. Was it increasing the radius timeout to 20 second or not, we didn't know that.