Container loses access to PIA (port-forwarding) intermittently and repeatedly: BindPort error / there has been a fatal error

[superkrups20056]

My container loses connection randomly.
I am running the latest version of docker and am on the :latest build
It searches for a new port after is loses connection and sometimes it can't find one leading to a fatal error.
I'm running PIA with port forwarding enabled. I use Synology.

The pertinent log is attached. Notice what happens after half an hour after I start my container.

Transmission copy.txt

[JB09]

Same issue, usually starts happening for me after the container has been running for over 24 hours.

the has been a fatal_error the port has been bound to 43540 Mon Nov 2 16:49:31 EST 2020 curl: (28) Connection timed out after 15001 milliseconds Mon Nov 2 17:10:23 EST 2020: bindPort error

[dalloro]

Same issue.

[zoidd]

same issue as well after 24 hours, restarting the container temporarily fixes it.

[ozhound]

I too have been having this issue for the last couple of days. The container doesnt crash it just starts reporting cant connect to trackers. (im using synology docker)

That just repeats over and over till i restart the container

I tried using the TCP package from PIA but i cant get it to stick on the container. i delete the ovpn files and copy the tcp ones over and they just get replaced when the container restarts. cant figure out how to force them to stick)

its pretty frustrating.

[oodport]

Latest docker on arm64 with PIA. Similar issue, however this is on a brand new container and it happens right away.
This occurs with all of the US configs I have tried (US New York, US New Jersey, US Chicago, US Florida, US Atlanta), HOWEVER, the container is able to connect when I use a European config (UK London, France, Spain), so this maybe this is a PIA issue.

[oodport]

It's probably my ISP (WoW) blocking the PIA nextgen US servers..

[ozhound]

I can connect fine to Sweden, Romania and Switzerland. (I've not tried any stateside end points)
it just looses connectivity anywhere between 1-60 minutes.

[steve-spiteri]

I face a similar issue using PIA and having the container running for an extended period of time.

You can see in openvpn/pia/update-port.sh the following in the bind_port () function:

  if [ "$(echo $pf_bind | jq -r .status)" = "OK" ]; then
    echo "the port has been bound to $pf_port  $(date)"		
  else  
    echo "$(date): bindPort error"
    echo $pf_bind
    echo "the has been a fatal_error"
  fi

Judging from the blank line in the log you can see that the pf_bind variable appears empty. I am still trying to narrow down what is causing the error in the pf_bind variable assignment.

I was having another issue from the container inexplicably creating a second tunnel interface which caused a bunch of problems with the health check script. #1462
It might be prudent to try some of the troubleshooting from there to see if you also have a second tunnel interface.

pf_host=$(ip route | head -1 | grep tun | awk '{ print $3 }')

I am wondering if this is related and in the pf_host variable assignment it is getting the wrong tunnel interface. I will attempt to validate.

[ozhound]

I checked if there where 2 TUN interfaces and there is not. so that rules that out

[haugene]

Hey. We've had some stability issues, and PIA has had some too. I would appreciate it if you could test with port-forwarding and health checks disabled. If that runs and is stable then you can try to activate them one by one and see when instability arises. Disable port forwarding by adding the environment variable:

DISABLE_PORT_UPDATER=yes

To disable health checks add --no-healthcheck to your Docker run command or in docker-compose:

healthcheck:
  disable: true

As for the RESOLVE: Cannot resolve host address error, that's DNS issue. Try to either specify the --dns flags as described in the documentation. Or add:

OVERRIDE_DNS_1=8.8.8.8
OVERRIDE_DNS_2=8.8.4.4

[reprz]

I´m also having this fatal error issue at the moment. I´m testing now with "--no-healthcheck" enabled. Let´s see what happens.

[reprz]

So far I´ve been running this with "--no-healthcheck" enabled for 2 hours without errors. As with healthcheck ON the fatal error could happen even 10 minutes after starting container.

[haugene]

@reprz Thanks for letting us know. A PR for a health check fix is in the making #1459

[superkrups20056]

Hi guys, I appreciate that my issue received so much traction. I resolved this issue by reverting back to 2.14. 3.0 is a hot mess right now! :)

[haugene]

Haha, don't blame you 😅 However I do believe that the last version is better now 🤞
Give it another try, or wait until 2.14 stops working I guess. Close this for now anyhow?