[inetquestion]

Linksys WRT3200ACM - DD-WRT r53339


For reason's I've not been able to capture, one or both wifi radios sometimes do not start after a reboot. On android phones there will be an exclamation next to the wifi icon indicating there is some internet connectivity problem. Other devices connect to another wifi AP with no issues.

Today I've found one of the radios (wlan0) in this weird state, where the channel, power, and type entries are different than when it's working normally (details below). After restarting the device with no other changes made, things typically return to normal.

Finding the root cause would be optimal, however there are no obvious errors in /var/log/messages as to why it failed or that it failed at all. For the meantime, automating finding and fixing appears possible. Is this the best/correct way to restart a wifi radio? Details on the proper way appear to vary widely over time...


service radio_off start
service radio_on start

clear; iw dev wlan0 info

###################
### Non-working ###
###################
Interface wlan0
ifindex 10
wdev 0x2
addr 00:00:00:00:00:00
type managed
wiphy 0
txpower 24.00 dBm

###############
### Working ###
###############
Interface wlan0
ifindex 10
wdev 0x2
addr 00:00:00:00:00:00
ssid 000000000000
type AP
wiphy 0
channel 157
txpower 5.00 dBm

[dale_gribble39]

Response in no particular order, save and except for point 1:

1. This is not an advanced networking subject.
   
2. Amplifying information as to which radio this is about.
   
3. The 'second' 5GHz radio is for DFS radar detection and has no antenna.
   
4. The only radios to worry about are the ones using the 88W8964 chip, not the 88W8887 chip.
   
5. It has always been advised both here and in the OpenWRT community to disable the 88W8887 radio.

https://wikidevi.wi-cat.ru/Linksys_WRT3200ACM

https://openwrt.org/toh/hwdata/linksys/linksys_wrt3200acm

https://openwrt.org/toh/linksys/wrt_ac_series

[inetquestion]

Seems there is no way to tell why radio is malfunctioning? Keep restarting until it works...

the 2nd 5ghz radio was already disabled; not in question.

[dale_gribble39]

Screenshots to prove the validity of your statement as wlan0 would be the 2.4 band if the interface renaming was straight across the board from athX to wlanX: Since your "working" example states "wlan0" but shows a 5GHz band channel, your information is suspect. EDIT: But I could be in error here, as "wlan2" should be the DFS radio. Either way, we need screenshots to see if your configuration is in error.

[egc]

Moved to the appropriate forum

[kernel-panic69]

There are only two reasons why the wifi radios will not start. Hardware failure due to lack of adding a cooling fan / environmental influence (electrical storm/surge) or bad / corrupt configuration. Please post screenshots of your wifi settings for review, thanks. It would also help to post output of 'cat /var/log/messages' from cold boot - especially with 'nvram set console_debug=1 && nvram commit' completed beforehand. Of course, we'll probably just dick around in yet another thread started by you without any solution achieved, so I'm probably wasting keystrokes here.

EDIT: A hard reset and rebuild configs from scratch is probably warranted given recent thread content.

[inetquestion]

Script running every hour to check if wifi was showing working state by running: "iw wlan0 info". If found in a bad state, all info is logged. While in a good state, only channel is logged. During this time, power went out multiple times. Caught a 3+ hour problem. This does align with your theory for dirty restarts. However this also happens on occasion from a restart initiated via cron when no other changes have been made to the system. A reboot fixes this every time. I'll add the additional information you've asked for as time permits. No reason to over speculate or assume a problem can only be solved one way. :)

While knowing the root cause is optimal, detecting the problem and being able to automate corrective actions may take precedent over a needlessly long back and forth. No offense intended toward anyone; it is what it is. Time will tell.




[code]
2023.08.07-17.00.00,channel 157 (5785 MHz), width: 80 MHz, center1: 5775 MHz
2023.08.07-18.00.00,channel 157 (5785 MHz), width: 80 MHz, center1: 5775 MHz
2023.08.07-19.00.00,channel 157 (5785 MHz), width: 80 MHz, center1: 5775 MHz
2023.08.07-20.00.00,channel 157 (5785 MHz), width: 80 MHz, center1: 5775 MHz
2023.08.07-21.00.00,channel 157 (5785 MHz), width: 80 MHz, center1: 5775 MHz
2023.08.07-22.00.00,channel 157 (5785 MHz), width: 80 MHz, center1: 5775 MHz
2023.08.07-23.00.00,n/a
23:00:00 up 39 min, load average: 0.00, 0.00, 0.00
Interface wlan0
ifindex 10
wdev 0x2
addr 00:00:00:00:00:xx
type managed
wiphy 0
txpower 24.00 dBm

2023.08.08-00.00.00,n/a
00:00:00 up 1:39, load average: 0.00, 0.00, 0.00
Interface wlan0
ifindex 10
wdev 0x2
addr 00:00:00:00:00:xx
type managed
wiphy 0
txpower 24.00 dBm

2023.08.08-01.00.00,n/a
01:00:00 up 2:39, load average: 0.00, 0.00, 0.00
Interface wlan0
ifindex 10
wdev 0x2
addr 00:00:00:00:00:xx
type managed
wiphy 0
txpower 24.00 dBm

2023.08.08-02.00.00,n/a
02:00:00 up 3:39, load average: 0.00, 0.00, 0.00
Interface wlan0
ifindex 10
wdev 0x2
addr 00:00:00:00:00:xx
type managed
wiphy 0
txpower 24.00 dBm

2023.08.08-03.00.00,channel 157 (5785 MHz), width: 80 MHz, center1: 5775 MHz
2023.08.08-04.00.00,channel 157 (5785 MHz), width: 80 MHz, center1: 5775 MHz
2023.08.08-05.00.00,channel 157 (5785 MHz), width: 80 MHz, center1: 5775 MHz
2023.08.08-06.00.00,channel 157 (5785 MHz), width: 80 MHz, center1: 5775 MHz
2023.08.08-07.00.00,channel 157 (5785 MHz), width: 80 MHz, center1: 5775 MHz
[/code]

[kernel-panic69]

First world problem, self-induced, nothing more. Otherwise, this thread would be filled with corroboration. No offense intended

[inetquestion]

Ah yes, the ole self-induced power outage. Haha...

Get off your high horse dude. If you have a suggestion which produces a better result than using a feature of the product (schedule reboot), which sometimes produces ill results I'm all ears. The script didn't cause the problem, it's being used to troubleshoot it. If you find that crazy also, i'm sorry.


TIA

[inetquestion]

Checks health of wlan0 & wlan1. Restarts afflicted radio if viable 'channel' is not found. Enable jffs support, then create cron job as follows:

### Runs once per hour
0 * * * * root /jffs/./dd-wrt_wifi_monitor.sh

[dale_gribble39]

And still no screenshots of applicable configuration information, complete system logs, or applicable output of nvram show, etc., but you still want help.

Forum guidelines and pointers, read first!

Official Forum Guidelines...How to get help the right way

Also, you have to reboot after setting console_debug=1, don't you?

[inetquestion]

Stopped asking for help when replies asking for information that's not needed as well as presumptuous and sarcastic replies were all that came. Why would you need screenshots of wifi config when an unexpected power cycle or problem due to rebooting is what happened and another reboot fixes it? It's clearly not a config problem.

Solved it with the aforementioned script; finds and fixes problem within 12 seconds.

No interest in arguing with the peanut gallery... because you're of course right, despite not being helpful.

Be well and have fun!

[Gameman Advanced Kid]

I know I am late responding to this thread, I wanted to get back to it as soon as I got home from work but I forgot about it\ until now.