Docker containers become unreachable after 20 minutes

[JNCK]

Hi!

 

I have issues with Docker containers becoming unreachable. I have multiple containers that are connected to a passthroughvpn container to route the traffic through my VPN. After 20 minutes or so all of the containers that use the --net=container:passthroughvpn parameter become unreachable.

 

In the system logs I see the following happening at that exact moment:

Mar 16 14:01:34 Tower kernel: docker0: port 1(vethacf7621) entered disabled state
Mar 16 14:01:34 Tower kernel: veth2a56a85: renamed from eth0
Mar 16 14:01:34 Tower avahi-daemon[6570]: Interface vethacf7621.IPv6 no longer relevant for mDNS.
Mar 16 14:01:34 Tower avahi-daemon[6570]: Leaving mDNS multicast group on interface vethacf7621.IPv6 with address fe80::c4cf:59ff:fe78:fc30.
Mar 16 14:01:34 Tower kernel: docker0: port 1(vethacf7621) entered disabled state
Mar 16 14:01:34 Tower kernel: device vethacf7621 left promiscuous mode
Mar 16 14:01:34 Tower kernel: docker0: port 1(vethacf7621) entered disabled state
Mar 16 14:01:34 Tower avahi-daemon[6570]: Withdrawing address record for fe80::ffff:59ff:fe78:fc30 on vethacf7621.
Mar 16 14:01:34 Tower kernel: docker0: port 1(vethd3ac1eb) entered blocking state
Mar 16 14:01:34 Tower kernel: docker0: port 1(vethd3ac1eb) entered disabled state
Mar 16 14:01:34 Tower kernel: device vethd3ac1eb entered promiscuous mode
Mar 16 14:01:34 Tower kernel: docker0: port 1(vethd3ac1eb) entered blocking state
Mar 16 14:01:34 Tower kernel: docker0: port 1(vethd3ac1eb) entered forwarding state
Mar 16 14:01:35 Tower kernel: docker0: port 1(vethd3ac1eb) entered disabled state
Mar 16 14:01:35 Tower kernel: eth0: renamed from vethc37dad7
Mar 16 14:01:35 Tower kernel: IPv6: ADDRCONF(NETDEV_CHANGE): vethd3ac1eb: link becomes ready
Mar 16 14:01:35 Tower kernel: docker0: port 1(vethd3ac1eb) entered blocking state
Mar 16 14:01:35 Tower kernel: docker0: port 1(vethd3ac1eb) entered forwarding state
Mar 16 14:01:36 Tower avahi-daemon[6570]: Joining mDNS multicast group on interface vethd3ac1eb.IPv6 with address fe80::e040:1bff:feac:bbcd.
Mar 16 14:01:36 Tower avahi-daemon[6570]: New relevant interface vethd3ac1eb.IPv6 for mDNS.
Mar 16 14:01:36 Tower avahi-daemon[6570]: Registering new address record for fe80::ffff:1bff:feac:bbcd on vethd3ac1eb.*.

 

A Docker container that is using --net=container:passthroughvpn becomes reachable again when I restart that specific container. After again around 20 minutes the container isn't reachable anymore.

 

I've searched online and I scrubbed my Docker and also increased the vDisk size to 50 GB for Docker. Still this issue persists.

 

Maybe someone knows what the above system logs mean?

 

Thanks!

tower-diagnostics-20210316-2207.zip

[JNCK]

Anyone who may be able to help me with this issue? Thanks!!

[Squid]

Since you're probably routing it through a VPN app from Binhex, maybe ask in his various support threads, or read his FAQ about that.,

[JNCK]

Since you're probably routing it through a VPN app from Binhex, maybe ask in his various support threads, or read his FAQ about that.,

I’m routing it through passtroughvpn of [mention]Dyonr [/mention]. I’ll post this in his support thread.

[Meldrak]

Hi,

 

Maybe same issue for me, all docker change network about 40 times per day, so docker network is lost for about 3 seconds

 

Aug 29 16:38:48 MNAS kernel: eth0: renamed from vethc37474d
Aug 29 16:38:48 MNAS kernel: IPv6: ADDRCONF(NETDEV_CHANGE): vethe47692b: link becomes ready
Aug 29 16:38:48 MNAS kernel: docker0: port 1(vethe47692b) entered blocking state
Aug 29 16:38:48 MNAS kernel: docker0: port 1(vethe47692b) entered forwarding state
Aug 29 16:39:02 MNAS kernel: docker0: port 1(vethe47692b) entered disabled state

 

in some docker I can see errors too

 

2021/08/29 16:38:49.681624 [info] error handling UDP packet: dns: buffer size too small
2021/08/29 16:38:52.682049 [info] error handling UDP packet: dns: buffer size too small

 

this issue appear 2 weeks ago and don't find anything to resolve it