The web-portal login page systematically fails to display during peak time. This article contains information regarding what data to collect and provide to your technical support provider if you find yourself in this situation. This data is relevant only if it is collected when the HTTP daemon is stuck: no login page is displayed at the very moment when the data is collected.  This article also presents a workaround to restart the HTTP daemon from the CLI.

If you observe that the web-portal login page systematically fails to display during peak time, the HTTP daemon may be stuck. One symptom of this issue is that no login page is displayed at the very moment when the data is collected.

An effective workaround is to restart the HTTP daemon from the CLI using one of the following methods:

# stop httpd
--wait 5 seconds--
# start httpd


OR

# test dump httpd

Open a terminal session to the affected WLC, log the terminal session in a text file and use the following commands in quick succession.

# set length 0

# set trace httpd level 10
# set trace stat level 10
# set trace ipc level 10
# set trace threadpool level 10


# show timedate
# show tcp
# show socket
# show proc x 3
# show load cpu history minute
# show load cpu history hour
# show load cpu history days
# show sessions
# show stat counters
# show ipc
# test dump httpd
# show tech
# show log buffer
# show log trace

# clear trace all

Provide the terminal log file and the HTTPD core file which was created using # test dump httpd to your JTAC contact or your technical support representative.