cross-posted from: https://leminal.space/post/41927223

I'm noob at Linux (it just works), so I have no idea how to start troubleshooting this weird issue, any help appreciated!

The setup: stock EndeavourOS, ProtonVPN, qBitTorrent bound to the VPN interface (and Transmission);

the issue: if qBit is running, no matter the load or the server (I have access to P2P-enabled servers), it kills my VPN connection in a matter of minutes, Transmission does not. It is reproducible, launching qBit means getting disconnected, without fail;

the timeline: no issues since I installed EndeavourOS (a year and a half or so), it started happening five to seven days ago, though since a few months ago I would get the same disconnection while qBitting, but only once a day and only while I was away and the system was locked, after I came back to it and reconnect, stable until shutdown. In the months between I haven't made any changes to my system, just got a new GPU a year ago;

what I tried: I initially though it was a VPN app problem, so I tried reinstalling it, both native and flatpak, after deleting all of its data of course.

Transmission works but I prefer qBit so I'd love to go back to using it.

Where do I start troubleshooting?

I currently can't be at my PC, so I can't troubleshoot "live", but I figured posting this early means more people seeing it and getting more input.

Thanks in advance!

ip link (VPN, no qBit): https://pastebin.com/RtwDqiEN

ip link (VPN, qBit): same identical output (verified with Kate, EndeavourOS text file reader)

journalctl qbit: https://pastebin.com/e2XCgGKy

journalctl PVPN: https://pastebin.com/DiPBvP5W

dmesg: empty for both "torrent" and "proton"

sudo dmesg -l err: virt/tdx: TDX not supported by the host platform

FINGERS CROSSED but it seems that lowering max connections DID fix it, though I'm currently on >20/10/5/4 and it's limiting. I'll try upping and see where it lands on, and hopefully that'll be it.

As per this comment

https://leminal.space/post/41934365/26536143

Proton must have changed something on their end since I used to run qBit with 500 max connections up >until the problem started

you are viewing a single comment's thread
view the rest of the comments
[–] 6 points 1 day ago (5 children)

I mentioned the starting paused thing because it lets you set the port before it starts talking to all the trackers, I think changing it shortly after starting it normally might trigger it reconnecting to all of them which also causes a lot of traffic. Currently I have it capped at 85 global connections

  • source
  • parent
  • hideshow 5 child comments
  • [–] [S] 5 points 1 day ago* (4 children)

    I've been on 20 max connections for the past 20-25 minutes and you were right, that seems to have fixed it! They must have done something on the backend because I used to run qBit with 500 max connections up until a few days ago with no issues. I'll try upping max connections a bit and see where I land on. Thanks! No idea why it didn't work on the first try

  • source
  • parent
  • hideshow 4 child comments
  • [–] 4 points 1 day ago (3 children)

    It’s been like this for quite some time for me, I want to say over a year, could be a policy that’s not applied everywhere or something though.

  • source
  • parent
  • hideshow 3 child comments
  • [–] [S] 3 points 1 day ago* (2 children)

    Damn. I mean, I still almost saturate my upload with just 20 connections, but I only seed one thing at a time. I'd rather seed more stuff, at lower speeds, but if this is it, then this is it. Thanks again, though, qBit works again!

  • source
  • parent
  • hideshow 2 child comments