Router disconnects all devices while Teams call (OpenWrt 21.02)

Hi there,

this topic is a follow-up on that one, where I mistakenly thought USB logging wasn't working:

Now I got working, USB drive-based logging, I got on with trying to make a Teams call, since I want to find out the origin of this rather serious bug.

The hardware: Linksys EA4500 v1
The software: OpenWRT 21.02 r16279.
The configuration: mostly stock. No updates applied, logs are supposed to be debug-level. rsyslog installed, logs redirected to USB-drive, buffer size increased to 512KB. Wifi access point turned on to allow one computer to connect. Not tried with Ethernet due to layout constraint, but the result was identical in previous trials.

Detailed results:

  • First call:

outside calling home: 22:19. Disconnected (In Windows, little globe appearing in the task bar instead of the Wifi symbol, any web page ending in 404 error).
Reconnected 22:22.

/mnt/sda1/log/system.log covering the first trial (copy-pasted from terminal)

  • Second call:

home calling outside: 22:25. Disconnected 22:26, same symptoms.
Reconnected 22:31.

  • Third call, not picked up:

home calling outside: 22:31. Disconnected 22:32.

System log covering the last two trials (from LuCI web interface):

system.log from /mnt/sda1/log/system.log covering the last two trials (copy-pasted from terminal):

The result: 100% failure rate.
A call made from Teams to another, distant Teams disconnects the computer behind the OpenWRT-running router (and probably all other devices attached to the same router, as previous trials showed). Nor the exact version of Teams nor the OS it's running on have any influence on the outcome.

What is really happening from these logs? More precisely, which component of OpenWRT is misbehaving?

I'd suggest that you start by disabing odhcpd service unless you plan to use IPv6

1 Like

Why would IPv6 cause problems?

I'm not sure how to disable the service without breaking something else. As there are no plans from local ISPs to roll out IPv6 to end-users, I guess I can safely turn off IPv6.

The odhcpd and lan sections from the buggy OpenWRT 21.02 installation show:

config odhcpd 'odhcpd'
        option maindhcp '0'
        option leasefile '/tmp/hosts/odhcpd'
        option leasetrigger '/usr/sbin/odhcpd-update'
        option loglevel '4'

config dhcp 'lan'
        option interface 'lan'
        option start '100'
        option limit '150'
        option leasetime '12h'
        option dhcpv4 'server'
        option dhcpv6 'server'
        option ra 'server'
        list ra_flags 'managed-config'
        list ra_flags 'other-config'

Whereas from a working OpenWRT 19.07 I get:

config odhcpd 'odhcpd'
        option maindhcp '0'
        option leasefile '/tmp/hosts/odhcpd'
        option leasetrigger '/usr/sbin/odhcpd-update'
        option loglevel '4'

config DHCP 'lan'
        option interface 'lan'
		option start '2'
		option limit '248'
		option leasetime '12h'
		option dhcpv6 'disabled'
		option ra 'disabled'
		option ra_management '2'

In both cases, the odhcpd section is the same. Which option would need to be modified in /etc/config/dhcp, and how would I achieve that from the GUI? The documentation isn't clear on that.

In LuCI go to Network>Interfaces then edit LAN interface, then go to DHCP server tab and set to disabled RA service, DHCPv6 and proxy NDP.

1 Like

Like this?

1 Like

Yes then restart the LAN interface.

@Cubytus Did changing the IPv6 settings resolve your issue?

Hi @rockjob , unfortunately not. Since I had to have a working LAN, I reverted to an older WD MyNet. Slower, but doesn't exhibit this crippling bug.

I still have the EA4500 powered on, though disconnected from the network, in case I want to test it again.

In the meantime, I was given a TP-Link Archer A6, so probably will be using it and retire the WD MyNet.

So, an unsolved bug :frowning: