Troubleshooting Tip: Data to be captured for WIFI performance issue troubleshooting
Description
Data collection should be performed during both off-peak and peak hours in order to obtain a comparison.
Solution
On the FortiGate GUI, take screenshots of Wireless Health -> Client count over time (screenshots of 1h, 1 day, and 30 days must be captured).
On the FortiGate GUI, locate the test PC -> Client monitor, filter on 'device' using the test PC’s MAC address. Take a screenshot of this output.
On the GUI, select 'Managed FortiAPs' and make sure the 'join-time' column is selected. Take a screenshot.
Capture the following debug information (output must be logged to a file):
diagnose wireless-controller wlac -d vap
diagnose wireless-controller wlac -c wtp
diagnose wireless-controller wlac -d sta
The live debug will start from now:
Set the admin timeout to 480 minutes to avoid unexpected Putty session time-outs. This can be reverted to defaults after capturing data.
- On the FortiAP:
- Open a first telnet session to the FortiAP where the test PC is connected, and capture dmesg on the FortiAP session must be logged to a file (dmesgbeforepeak.log and dmesgdureinpeak.log).
cw_diag repeat 10000 1 "dmesg -c"
This command will continuously capture the dmesg on the FortiAP, and a prompt will be returned after 10000 counts. If there is any ping lost, it should be captured on this output.
- At the same time, open another telnet session to the FortiAP and capture cw_debug on the FortiAP. The session must be logged to a file (cwdebug.log before and during peak hours).
FAP# # cw_debug on
FAP# # cw_debug app cwWtpd 0x7fff
-
Air sniffer.
At the same time, a wireless sniffer must be taken near the test PC. Make sure there is NO filter on the air sniffer. The channel must be set to the one used by the test PC (association to FortiAP).
- On the FortiGate.
At the same time, on the FortiGate, the following information must be captured:
- debug cw_acd on the FortiGate from SSH.
- Sniffer trace on the FortiGate from SSH - session must be logged to a file.
To filter the traffic of FortiAP where the test PC is connected, the following must be done (this is an example) using:
diagnose wireless-controller wlac -c ws
Find the FortiAP WTP session of the FortiAP used by the test PC.
This is the output taken from a lab as an example:
-------------------------------WTP SESSION 2----------------------------
WTP session : 0-192.168.222.82:5246 CWAS_RUN ïƒ impacted FAP session
indev : 4(wan2)
in_ifIdx : 4(wan2)
mesh uplink : ethernet
id : FP221B3X11000009
Using the output above, use the following diagnose command to turn on the filter:
diagnose wireless-controller wlac wtp_filter FP221B3X11000009 0-192.168.222.82:5246 4
Then the following debug can be enabled
diagnose debug console timestamp enable
diagnose debug application cw_acd 0x7fff
diagnose debug enable
Note: As soon as the above commands are enabled, lots of logs will come up, but it will be possible to copy and paste the below sniffer commands in the same window without any error.
Sniffer trace must be captured at the same time on the same SSH session:
diagnose sniffer packet any "host " 6 0 a
The output must be captured to a file (FGT.log before and during peak).
- Collect ping test results on STA (test PC).
While ping is running, take a screenshot whenever there is a ping loss and paste it into a Word file with the file name as sta_ping_log. Make sure ping timeouts and laptop time are visible on screenshots.
Leave the debug (1., 2., 3., and 4.) running during the time that the problem is present. The duration of the capture will depend on the presence of the problem (ping loss).
To troubleshoot this issue, capturing during the problem state only will be helpful.
To analyze the data, open a support ticket and attach the zipped data along with the latest config file.