wifi dies every 4 days, requires reboot -- acsd start up at exactly 11pm?

[jorhett]

I'm seeing a consistent problem where the router becomes inaccessible every 4 days after 11pm PDT. Physical reboot via power switch is required. I can't get to the console to see the logs, so I sent up syslog and sent all logging to a local node and started looking at logs to try and figure out what's going on. I now have 5 instances of this happening with some clarity. It would appear to be a cron job causing the failure, from what I see in the logs each time this happens:

<15>1 2024-06-25T23:07:44-07:00 bruinen-6969758-C rc_service - - - rc_service: ntpd_synced 2074:notify_rc restart_diskmon
<15>1 2024-06-25T23:07:44-07:00 bruinen-6969758-C disk_monitor - - - disk_monitor: Finish
<15>1 2024-06-21T23:00:02-07:00 bruinen-6969758-C acsd - - - acsd: eth7: Selecting 5g band ACS policy
<15>1 2024-06-21T23:00:02-07:00 bruinen-6969758-C acsd - - - acsd: acs_init_run(1111): acs_start eth7 cannot select chanspec with the wrong info
<15>1 2024-06-21T23:00:05-07:00 bruinen-6969758-C acsd - - - acsd: eth7: Selecting 5g band ACS policy
<15>1 2024-06-21T23:00:05-07:00 bruinen-6969758-C acsd - - - acsd: acs_init_run(1111): acs_start eth7 cannot select chanspec with the wrong info
<15>1 2024-06-21T23:00:08-07:00 bruinen-6969758-C acsd - - - acsd: eth7: Selecting 5g band ACS policy
<15>1 2024-06-21T23:00:08-07:00 bruinen-6969758-C acsd - - - acsd: acs_init_run(1111): acs_start eth7 cannot select chanspec with the wrong info
<15>1 2024-06-21T23:00:11-07:00 bruinen-6969758-C acsd - - - acsd: eth7: Selecting 5g band ACS policy
<15>1 2024-06-21T23:00:11-07:00 bruinen-6969758-C acsd - - - acsd: acs_init_run(1111): acs_start eth7 cannot select chanspec with the wrong info

That last message is repeated thousands, tens of thousands, of times before the router becomes inaccessible and also stops syslogging.

RT-AX86U running 3004.388.7. All wifi channels are fixed. After a reboot, acsd is not running. So it appears something in the cron jobs is starting up acsd, which is killing off the router. That said, there's nothing in crontabs that relates

myname@bruinen:/tmp/home/root# ls /var/spool/cron/crontabs/
myname

myname@bruinen:/tmp/home/root# crontab -l
50 0 */7 * * service restart_letsencrypt #LetsEncrypt#

I've poked around in the OS and I can't figure out where this is being run, nor can I find how disk_monitor is being invoked, either, even though it's just before.

 

[Adooni]

https://www.snbforums.com/threads/a...ot-select-chanspec-with-the-wrong-info.85364/

 

[jorhett]

https://www.snbforums.com/threads/a...ot-select-chanspec-with-the-wrong-info.85364/

I'm going to guess that you're suggesting I track this existing/unresolved thread? Because I don't see any answer there.

I can confirm my configuration is much like the others: AX86U with single AC86U aimesh node. People are reporting that "Internet goes down" but I can't confirm that--but I lose access over Ethernet as well. It seems the router itself stops responding.

For example, if I try to SSH or HTTPS to the router I will get a TCP SYN ACK but no other communication. So the IP stack is operating but the OS and/or services that should reply have failed/overloaded/etc.