Closed illuvattarr closed 1 year ago
WIREGUARD_IMPLEMENTATION=userspace
(pull the latest image, that was done yesterday) see if it helps?Honestly I doubt this has to do with the vpn protocol. Maybe Surfshark filters on their wireguard vpn servers, but anyway that's out of our control.
Why using ` depends_on:
and
network_mode: container:gluetun-wireguardfor your torrent instead of
network_mode: "service:vpn"`? Thanks for the suggestions. For the few days I tried with openvpn, it never happened at least. And with wireguard, it happened again immediately.
I added WIREGUARD_IMPLEMENTATION=userspace
and also replaced depends_on: - vpn
and network_mode: container:gluetun-wireguard
with network_mode: "service:vpn"
in the compose code. And so far so good at least; wireguard is working without issues at the moment.
What does the wireguard_implementation variable add?
Well it just started happening again after all. All torrents stay stuck on finding metadata.
I tried to switch to Deluge but there torrents stay stuck on downloading 0.00% with different tracker status errors; skipping tracker announce (unreachable)
or host not found (authoritative)
.
Im using the compose code below to set deluge up:
version: "3.5"
services:
vpn:
image: qmcgaw/gluetun:latest
container_name: gluetun-wireguard
cap_add:
- net_admin
environment:
- VPN_SERVICE_PROVIDER=surfshark
- VPN_TYPE=wireguard
- WIREGUARD_PRIVATE_KEY=key
- WIREGUARD_ADDRESSES=10.14.0.2/16
- SERVER_COUNTRIES=Netherlands
ports:
- 8080:8112
- 6881:6881/tcp
- 6881:6881/udp
deluge:
image: lscr.io/linuxserver/deluge:latest
container_name: deluge
network_mode: "service:vpn"
environment:
- PUID=1027
- PGID=1000
- TZ=Europe/Amsterdam
- DELUGE_LOGLEVEL=error #optional
volumes:
- /volume1/docker/deluge:/config
- /volume1/video/downloads:/video/downloads
restart: unless-stopped
I would try to repro using a different VPN provider to rule out the provider being the issue.
@illuvattarr
There's a mistake in both your stacks. The linuxserver/deluge and linuxserver/qbittorrent containers both need a volume binding to /downloads
not /video/downloads
It turns out it the Deluge errors were due to another mistake in the compose code; PGID for me is 100 and not 1000 and I forgot to change it when copying the default compose code from the Deluge docker website... So far Deluge is working now without issues like torrents staying stuck after changing 1000 to 100.
What does the wireguard_implementation variable add?
The kernelspace implementation (written in C in your Linux Kernel) is meant to be faster (at least as fast) as the userspace implementation (written in Go, built in Gluetun). So previously and now by default, Gluetun picks the kernel one if available and falls back on the userspace one (for older/custom kernels). Now you can choose, really to debug and bench testing.
So far Deluge is working now Sometimes I can kickstart the torrents by changing the port in the settings of qbittorrent webui, or changing it back again.
Cool, I'll close this issue for now, feel free to comment if it dies again.
Maybe, maybe... it's due to how Qbittorrent handles IP changes/VPN reconnections ([healthcheck] program has been unhealthy for 11s: restarting VPN
which can happen occasionally), and Deluge may handle it better. I personally use Deluge but I have not really a preference on this, so I can't say more :wink:
Is this urgent?
No
Host OS
docker synology NAS
CPU arch
None
VPN service provider
Surfshark
What are you using to run the container
Portainer
What is the version of Gluetun
latest image from qmcgaw/gluetun
What's the problem π€
I'm running gluetun together with qbittorrent in docker through compose in portainer. I'm trying to use wireguard from VPN provider Surfhshark, and often the torrents/magnet links stay stuck on finding metadata. Using openvpn does work, but I'd prefer using wireguard since speeds are higher with wireguard.
See the configuration for the compose code I use to create the stack in portainer.
Sometimes I can kickstart the torrents by changing the port in the settings of qbittorrent webui, or changing it back again. But after a while, they stay stuck again on finding metadata.
Share your logs
Share your configuration