Some strange observations, having 3 interfaces to supervise: WAN (eth1), WIFI-WAN, 3g-WWAN
Note, that I do not use ipv6 on my custom image, but iptables/ip6tables present. Because impossible to compile image with ipv6 disabled.
syscntrl.conf:
net.ipv6.conf.all.disable_ipv6 = 1
net.ipv6.conf.default.disable_ipv6 = 1
net.ipv6.conf.lo.disable_ipv6 = 1
net.ipv6.conf.default.forwarding=0
net.ipv6.conf.all.forwarding=0
Issue 1:
When having WAN and 3g-WWAN up and running, when
ifdown wan ; ifup wan then
Sat Oct 22 06:29:35 2022 user.warn mwan3-hotplug[4048]: failed to add 192.168.1.0/24 dev eth1 proto static scope link metric 10 to table 1
Sat Oct 22 06:29:35 2022 user.warn mwan3-hotplug[4048]: failed to add default via 192.168.1.1 dev eth1 proto static src 192.168.1.2 metric 10 to table 1
However, it looks like failover works, both for down and up. Any comments, what is gong on here ?
Issue 2:
When unplugging WAN (eth1) directly at my router, I see normal messages:
Sat Oct 22 06:36:27 2022 user.notice mwan3-hotplug[27050]: Execute ifdown event on interface wan (unknown)
Sat Oct 22 06:36:28 2022 user.info mwan3track[11208]: Detect ifdown event on interface wan (eth1)
Sat Oct 22 06:36:28 2022 user.notice mwan3track[11208]: Interface wan (eth1) is offline
and failover (to 3g-wwan) works.
Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref Use Iface
default 10.64.64.64 0.0.0.0 UG 50 0 0 3g-wwan
10.64.64.64 * 255.255.255.255 UH 0 0 0 3g-wwan
192.168.30.0 * 255.255.255.0 U 0 0 0 br-lan
However, reconnecting the WAN-cable again at my router causes a reboot, without any log messages. When doing all this, I have active web connections (i.e. mqtt)
This had no bad consequences on my 21.02.3, though .
And no problem, in case /etc/init.d/mwan3 stop/disable
The interesting part is, that disconnection/reconnection of other end of WAN-cable, at my ISPs router, has no negative effect. Normal switchover happens. So it looks like some "bad event" happens for mwan3, when eth1 removed/inserted directly at router.
Any idea, how to trace down this issue on 22.03.2, which is easily reproducable ?
Issue 3:
Is it possible to get rid of these redundant processes:
1670 root 1636 S {mwan3rtmon} /bin/sh /usr/sbin/mwan3rtmon ipv6
1799 root 1504 S ip -6 monitor route
No IPV6 in use.