I'm using the "O2 Box". I had no problems with 17.01.4. Now with 18.06.0 I have regular PPP hangups. The DSL connection seems to keep alive but even manual retry attempts fail.
(Sorry for the disabled links. Seems like some silly forum configuration disallows me to post links)
I tried to "reconnect" the WAN interface which failed.
I tried to disconnect the DSL wire in hope a complete "reinitialization cycle" will fix this but this failed.
A reboot fixes this but the last time this hangup happened was just 2 hours ago.
I tried to go back to lede-17.01.4 but the new firmware doesn't take my old firmware image for flashing.
Is this a known bug?
How to get back to my 17.01.4 version?
I have to disconnect the OpenWRT router for now and connect some proprietary one.
daemon.warn pppd: Timeout waiting for PADO packets
daemon.err pppd: Unable to complete PPPoE Discovery
is this a known bug?
You can find multipe posts on failed Lantiq+PPPoE reconnects (pppoe-reconnection-failing-lantiq-built-in-modem). I'm not sure myself, but by checking out this commit openwrt/openwrt#555c5923 and its references higher tolerances with the LCP (link control protocol) echo might help to not disconnect in the first place by a missing echo: adjust lcp-echo-failure/lcp-echo-interval from their 5/1 default values fit for your ISP connection (5/5 helped one of the commentors) or set them to 0 in /etc/ppp/options to disable the echo failure modes. If you use a snapshot from after 01/2019 with the commit, passing option keepalive 0 in the config interface 'wan' section will disable the echos.
From my observation, disconnects happen when there's lots of throughput and 5 echos can go missing within 5 seconds. Reboots always helped to resolve the reconnection issue too.