[Bug]: Constantly logging out of Navidrome server

[loboda4450]

I have already checked through the existing bug reports and found no duplicates

• Yes

App Version

1.3.0

Music Server and Version

Navidrome v0.59.0

What local environments are you seeing the problem on?

Web Firefox

What happened?

Being constatnly logged out of my navidrome server even though playback continues to play. Wiping cache + cookies resolves problem until next browser start.

Steps to reproduce

1. Log in
2. [Optional] Play something
3. Wait
4. Be logged out :/
5. Logging in is now corrupted

Relevant log output

No such unfortunately

[JuniorJPDJ]

Same issue on same navidrome instance (other user), also using firefox.

[td31]

Also facing the same issue and also using Navidrome and Firefox.

[jeffvli]

If you can check what the first network request you're receiving a HTTP 401 on when you're getting invalidated that would be helpful.

Also, you can increase the ND_SESSIONTIMEOUT configuration to a larger value. If you leave your browser tab idle for long periods of time and then return to it without refreshing the page first, it may try to access your server using stale access tokens which causes the logout.

By default it is set to 48h, but you can try increasing it to 168h (1 week) or 336h (2 weeks).

[loboda4450]

For me, the very first one to fail is GET /api/album?_end=20&_order=DESC&_sort=random&_start=0&missing=false. I'll adjust the session timeout to 168h, but it's not the cause I believe. I was using this session earlier today and I am quite confident it didn't exceed the configured 24h threshold.

[loboda4450]

Yeah, it hasn't helped indeed, now when accessing the feishin the first to fail is GET /api/song/42901cccdd25ca0a3f57cedcb643d821. Looks like it's pretty random tbh, it fails with the first one operation to be done.

[JuniorJPDJ]

Check the headers sent to requests responding with 200 and those responding with 401 in a free time please. Maybe this will help.

[loboda4450]

I checked that before and nothing stung my eye, I'll check again later today to confirm.

[zeroquinc]

Getting the same issue on Firefox.

EDIT: I noticed everytime I got logged out, it added the server again, so I had the same server added like 6 times.

I changed - SERVER_LOCK=true to - SERVER_LOCK=false in my docker compose, removed all the servers and added it again, and now it doesn't seem to log me out anymore.

[jeffvli]

Check the headers sent to requests responding with 200 and those responding with 401 in a free time please. Maybe this will help.

Agreed with this.

Check the Request Headers and copy the X-Nd-Authorization header for one the requests that is 401ing.

Use a JWT decoder and check what the exp (expiry date) and iat (issued at date) values are for the token.

You can then use these values to determine if your Navidrome session configuration is working as expected or if somehow an incorrect token is being stored.

[td31]

Getting the same issue on Firefox.

EDIT: I noticed everytime I got logged out, it added the server again, so I had the same server added like 6 times.

I changed - SERVER_LOCK=true to - SERVER_LOCK=false in my docker compose, removed all the servers and added it again, and now it doesn't seem to log me out anymore.

Seeing the same issue with the duplicate servers. Followed the steps mentioned about updating the docker compose file then removing the duplicate servers and it seems to be working again without issue.

[loboda4450]

Indeed looks like it resolves the issue, however I will still test it. @JuniorJPDJ try that too.

[not-siblers]

This is highly irritating as I have to keep disabling and enabling the server lock setting. From what I can see the issue seems to be that Feishin is adding the server each time you log in, rather than using the stored server settings.

The ability to remove the list of saved servers on the login screen would solve this in the short term.

[CaptainKey3s]

I have the same issue. I also changed SERVER_LOCK=true to - SERVER_LOCK=false and deleted the duplicated servers. I still need to test, if the issue occurs again.