"Host not found (non-authoritative), try again later": how to disable DNS cache?

[lgrn]

qBittorrent & operating system versions

qBittorrent-nox v5.0.4 on Debian Trixie

What is the problem?

All torrents are silently failed (not errored) with all trackers showing error:

Host not found (non-authoritative), try again later

DNS is working fine, I've also tried changing my DNS server with resolvectl to another one that I also verified is working through dig, but qBittorrent is still doing nothing, regardless of reboots, removing and re-adding the tracker etc.

This issue seems to talk about DNS caching, with an unanswered question on how to disable it: #9045 (comment)

It's quite possible that the DNS resolution did fail at some point, perhaps through a race condition, but it's concering to me that it doesn't seem to retry it when DNS is working fine, and that it seems like nothing can be done to force a re-check.

So are there any ways to force a re-check, configure a lower timeout, or any other ideas on what might be causing this?

Steps to reproduce

Intermittent

Additional context

No response

Log(s) & preferences file(s)

No logs on STDOUT

[shama84]

DNS is at OS level, no QBT. The issue you mention points to a CDN wrong IP (Content Delivery Network) which is ou of your boundaries but what you can do is using a DNS servers more "reactive".

You local DNS cache should be cleard under Debian with (may need sudo) "resolvectl flush-caches". This will flush the local machine DNS cache and all new queries will go to the configured DNS server for resolution. However and most likely the configured DNS server is not very reactive and doesn't update fast enough and you will fall in the same rabbit hole.

Have you change you server DNS servers to something else than ISP ones? (Cloudflare, Cloud9, whatever is your flavor)?

[lgrn]

I probably should have mentioned that qbittorrent-nox is running in an LXC container, so that might affect things, but still:

• The issue persisted at least 15 minutes, likely longer
• During the entire time, DNS queries on the "OS level" (done with dig) worked fine.
• During the issue, I tried rebooting -- in reality more like restarting all processes since it's a container -- and I tried changing the DNS server with resolvectl, neither of which made any difference.

If it happens again I'll try flush-caches but it doesn't sound right to me; the exact DNS query that qbittorrent told me failed worked fine with dig, and I got a feeling (I have no proof) that restarting qbittorrent did not result in a new DNS query being performed, it just looked like the error was "sticky" from the last attempt.

[shama84]

Is Trixie in stable branch now? Last time I checked was Bookworm still the one to use (and deployed in LXC).
Maybe the root cause of issue if not and still in testing branch

The DNS servers to change is so inside the LXC container, did you change them there?
Another source of issue I've experienced is the way to change DNS in Debian
https://wiki.debian.org/NetworkConfiguration#Defining_the_.28DNS.29_Nameservers

the default was problematic for me and I used resolvconf way instead:
https://www.ionos.com/digitalguide/server/configuration/how-to-set-dns-on-debian/#:%7E:text=the%20DNS%20field.-,How%20to%20set%20your%20DNS%20server%20using%20the%20configuration%20file,-You%20can%20also