I'm using OpenWrt SNAPSHOT r25219-6e2962d4c5 for my GL-MT6000. I have multiple VLANs being used on both radios. I've noticed that both radios start randomly giving "Device is not active" and the APs on them start showing "Wireless is not associated". It seems to change and when I disable one AP another one will come online. I don't think I can use any of the wireless networks when this happens.
I did find that if I manually disable all of my 5GHz APs and restart my router, I am able to use all of the 2.5GHz ones just fine. If I manually enable all of the 5GHz ones after that reboot, I can use both radios just fine and I haven't run into any issues today since doing that. If I reboot, I go back to the same issue.
Disabling all 5GHz APs before and after rebooting is my current workaround, but is there a way to make it so this isn't happening? I removed all of the explicit MAC addresses for my lan devices, but left the explicit MAC for eth1 there, which is what let me get this working at all.
I am happy to share any configuration files if you need them, if this is something that needs to be fixed on OpenWrt's side. If it is something that I can configure, I would love to know if there's a way to do so.
You're not alone. I have the same oddball problems but I can't be clear as to what is happening and so I can't explain what is happening yet. Restarts and reboots don't fix it sometimes. It seems that I only have problems after flashing an update.
My fix (temp) when I can't get it to clear is...
rename /etc/config/wireless to anything. I use wireless.OFF
Then reboot router and let Openwrt create it's own wireless
now your radios are working but not configured.
Go and delete the new /etc/config/wireless and then rename your old renamed wireless.OFF back to wireless.
Now log out and log back in. Give it a few seconds and check your wireless network setup.
Your old original wireless config will be back and functioning.
Sounds crazy. Sounds stupid. Sounds like BS. Maybe so.
But mine will then run without fail until the next flash update.
This firmware is in beta and it is a pretty young version. base-files package is updated ALL the time. Gotta wander if that my have something to do with it. Alot of testing on drivers is going on right now too.
Once I get mine running it's just as solid as can be. I love it.
You can always delete wireless and reboot and then rebuild your wireless config from scratch step by step. Tedious but likely fix it too.
I run five vlans and a guest network with trunk lines to AP's. Sounds like our setups my be similar.
Unfortunately that doesn't do anything. Manually running wifi reconf after it boots also doesn't do anything. Manually disabling and restarting the radios and APs doesn't let me connect to any of the APs on either radio.
I'm not sure if these are related, but I see these errors in my system.log:
Sun Feb 18 21:35:38 2024 daemon.notice netifd: radio1 (5234): Device setup failed: HOSTAPD_START_FAILED
Sun Feb 18 21:36:08 2024 daemon.notice netifd: radio0 (5233): Command failed: ubus call hostapd config_set { "phy": "phy0", "config":"/var/run/hostapd-phy0.conf", "prev_config": "/var/run/hostapd-phy0.conf.prev"} (Request timed out)
Sun Feb 18 21:36:08 2024 daemon.notice netifd: radio0 (5233): Device setup failed: HOSTAPD_START_FAILED
Sun Feb 18 21:39:18 2024 daemon.notice netifd: radio0 (7232): Command failed: ubus call hostapd config_set { "phy": "phy0", "config": "", "prev_config": "/var/run/hostapd-phy0.conf" } (Request timed out)
The problem sounds like some timing problem / race in reading the 5 Ghz firmware blob in time before the radio should be initialised. The first boot after flashing may take longer that normally, as overlayfs is being created etc.
Likely nothing. Base-files is actually updated quite rarely. Its versioning just claims so, as it is based on the total commit count, not changes in the package itself.
Maybe it is a symptom of the race described here?:
If affected, then the phy will have incorrect mac address at boot. It will fix itself if you do anything causing the phys to be "hotplugged" again. Which matches the description.
In my case, I "fixed" the issue by disabling block mount, as this was causing delays making the phys lose the race every time. But that's an arbitrary fix. The underlying problem is still there, and can't be fixed unless we get the driver to configure the proper mac addresses
Since you use a strange ipaddr and a lower dns i want you to concentrate to your dhcp here:
you start for clients is already 100, but your dns shows 20 as a client, you should also be carefully with the range, since your ipaddr shows 192.168.15.32 as gateway ip, but your start is 192.168.15.100 for dhcp clients until 192.168.15.250, just make sure your clients dont go over 254
And then i see you specify the default route to gateway 192.168.15.25 but does this router exist?
And then i see multiple entries of this here:
eth1 is supposed to be the port serving wan, but lan1 - lan5 these are lan ports which includes one of the 2.5gb ports aswell, i suspect these entries are wrong.
I used auc to try to upgrade to the lastest SNAPSHOT version, and I think I broke my device. I tried reset and connectivity using LAN1 and WAN, but I don't get any where, the device flashes blue twice then reboots every 90s or so. I tried failsafe mode in the documentation below, but the device does not seem to change is behaviour. Anythin else I can try before opening it up and trying to get console??
If you've only ever flashed sysupgrade images, you can use GL.iNet's built-in uboot recovery environment to flash a working image by holding down the reset button for 30 seconds while powering on the router.