Homarr•16mo ago

502 Bad Gateway

Hello all, happy Homarr user here! 🙂 Beginning from today after the container has been updated to the latest version (v0.13.3) I've got 502 Bad Gateway if I try to open Homarr. I'm using it behind nginx as a subfolder . I investigated a bit, but I didn't find any obvious, until I spot the problem in Homarr log: "ready started server on 127.0.1.1:7575, url: http://127.0.1.1:7575" Do you see it? Not? Let me help: 127.0.1.1. Any idea what's the problem? How can I solve it? One remark: I'm using docker host networking.

64 Replies

Manicraft1001•16mo ago

Hi, thanks for your post. This is a known issue with 0.13.3. Do you have authentication enabled? @randomvalid

randomvalidOP•16mo ago

Hi! No, I don't have. Should I modify nginx to point to 127.0.1.1 instead of 127.0.0.1?

Manicraft1001•16mo ago

No. The IP you're seeing is only related to the container's internal network Can you post your docker run or compose?

randomvalidOP•16mo ago

Sure, here it is: sudo docker run --name homarr \ --network host\ -d \ --restart always \ -v /container/homarr/config:/app/data/configs \ -v /container/homarr/icon:/app/public/icons \ -d ghcr.io/ajnart/homarr:latest

Manicraft1001•16mo ago

Thank you. We'll look into it. FYI @tagaishi

randomvalidOP•16mo ago

Since I'm using host networking, the modification in nginx is actually working, so it can be a workaround others with the same setup - I've just tested it.

Manicraft1001•16mo ago

Wait, what modification did you make? @meierschlumpf @ajnart could this be, because we changed the base image? Wasn't this an issue with the base image for some reason?

randomvalidOP•16mo ago

Okay, so I have an nginx reverse proxy with a bunch of location directives, originally it pointed to 127.0.0.1, but after that I changed to 127.0.1.1 and it works. Original: location / { proxy_pass http://127.0.0.1:7575/; } Workaround: location / { proxy_pass http://127.0.1.1:7575/; } As I mentioned I'm using docker host networking, so maybe that's the reason why this workaround is working for me.

Manicraft1001•16mo ago

Okay, thanks for the insight. The previous version worked without that workaround, right?

randomvalidOP•16mo ago

Welcome! Yes that's right!

Manicraft1001•16mo ago

Okay, thanks for telling us We'll look into it, but I do not have an ETA

randomvalidOP•16mo ago

No problem, it is usable now, whenever you release the fix I will write my config back 🙂

Serenaphic•16mo ago

What if you replace 127.0.1.1 in nginx by localhost, does it work? Also, check your /etc/hosts to see if the localhost line was removed or changed somehow

randomvalidOP•16mo ago

/etc/hosts seems fine - 127.0.0.1 localhost, 127.0.1.1 {fqdn} What do you mean by that?

Serenaphic•16mo ago

Proxy_pass http://localhost:7575/;

randomvalidOP•16mo ago

Let me check... That works also

Serenaphic•16mo ago

There you go :) All 127.x.x.x are loopback IP's, so setting to localhost should cover it no matter what your fqdn does

randomvalidOP•16mo ago

Thank you! I understand, anyway it's a good tip 😉

Serenaphic•16mo ago

Btw the fqdn is set by nginx and the OS, homarr has nothing in place that would touch that somehow So I don't really know how that problem hasn't shown earlier

randomvalidOP•16mo ago

I know, but somehow homarr binds to that adress rather than 127.0.0.1

Serenaphic•16mo ago

Maybe before it was set as

127.0.0.1 localhost {fqdn}

127.0.0.1 localhost {fqdn}

instead

randomvalidOP•16mo ago

Yes that is possible Sorry, I've just realized I was wrong... it do not work. localhost resolves 127.0.0.1, the hostname or fqdn resolves 127.0.1.1. So if I use localhost in nginx that won't solve the problem. (I tested one instance, while modified another one... 😄 )

Serenaphic•16mo ago

Ah, that's too bad.

randomvalidOP•16mo ago

Host is debian bookworm, and this explains: https://serverfault.com/questions/363095/why-does-my-hostname-appear-with-the-address-127-0-1-1-rather-than-127-0-0-1-in

Server Fault

Why does my hostname appear with the address 127.0.1.1 rather than ...

This may be a bit of a noobish question, but I was taking a look at /etc/hosts on my new Xubuntu install and saw this: 127.0.0.1 localhost 127.0.1.1 myhostname On most 'nixes I've used, the secon...

Serenaphic•16mo ago

And if you were to return to 0.13.2, is that issue persisting?

randomvalidOP•16mo ago

I can check it, but I doubt. Issue is gone if I revert to 0.13.2.

Serenaphic•16mo ago

Alright then there is indeed a problem somehow from our side

randomvalidOP•16mo ago

I think also, should I report it as a bug on github?

salty•16mo ago

GitHub

0.13.3: 500 Internal Server Error on homepage when password is enab...

Environment Docker Version 0.13.3 Describe the problem Just upgraded to 0.13.3. I had password enabled and was unauthenticated and the homepage now displays Internal Server Error. Note that there i...

Serenaphic•16mo ago

Possibly Yes

randomvalidOP•16mo ago

I think these two are seperate issues, I don't have authentication, and I get back a bad gateway, homarr is working but on a wrong address...

Serenaphic•16mo ago

Although the problem isn't exactly the same, they only both stem from using reverse proxying, the cause might be the same

salty•16mo ago

Yeah, I just noticed the reverse proxy link. Could be wrong.

Meierschlumpf•16mo ago

Are you currently using image tag 0.13.3 or latest? Maybe if you change to 0.13.3 it works? (Idk why it should but maybe) Then docker would just have a stroke with the current image

randomvalidOP•16mo ago

I'm using the latest tag, but I think that is exactely the same, so latest is 0.13.3

salty•16mo ago

they point to the same hash on GHCR at least.

Meierschlumpf•16mo ago

yeah that's true (Does it rebuild if you get the same hash for another image name?)

salty•16mo ago

it will not re-download it, no.

Meierschlumpf•16mo ago

Okay I see

Serenaphic•16mo ago

@meierschlumpf the only thing I can see causing problem would be the next dependency update

salty•16mo ago

but clashing hashes are not likely

Meierschlumpf•16mo ago

It's unlikely in my opinion but of course it could be the case. I'm just wondering why nobody using dev had the problem then

randomvalidOP•16mo ago

Tried it out, using exact tag, issue is back...

Meierschlumpf•16mo ago

Ok thanks for your effort

Serenaphic•16mo ago

But it doesn't make sense because we just went from .10 to .19 so unlikely to me too

salty•16mo ago

The funny thing about the issue I linked is that you can make the page work if you rollback, login, then update to the new version and use the login session from earlier, so that is related to the login bits. 😄

Meierschlumpf•16mo ago

Is dev working? Because there should be exactly the same changes but it is another hash

randomvalidOP•16mo ago

Yeah but I guess, if you clear the session cookie you won't be able to login again...

salty•16mo ago

yeah, that was my point. login triggers this

Failed to proxy https://172.19.0.11:7575/en/login Error: write EPROTO 980B5FDB137F0000:error:0A00010B:SSL routines:ssl3_get_record:wrong version number:../deps/openssl/openssl/ssl/record/ssl3_record.c:354:

    at WriteWrap.onWriteComplete [as oncomplete] (node:internal/stream_base_commons:94:16) {

  errno: -71,

  code: 'EPROTO',

  syscall: 'write'

}

Error: write EPROTO 980B5FDB137F0000:error:0A00010B:SSL routines:ssl3_get_record:wrong version number:../deps/openssl/openssl/ssl/record/ssl3_record.c:354:

    at WriteWrap.onWriteComplete [as oncomplete] (node:internal/stream_base_commons:94:16) {

  errno: -71,

  code: 'EPROTO',

  syscall: 'write'

}

Failed to proxy https://172.19.0.11:7575/en/login Error: write EPROTO 980B5FDB137F0000:error:0A00010B:SSL routines:ssl3_get_record:wrong version number:../deps/openssl/openssl/ssl/record/ssl3_record.c:354:

    at WriteWrap.onWriteComplete [as oncomplete] (node:internal/stream_base_commons:94:16) {

  errno: -71,

  code: 'EPROTO',

  syscall: 'write'

}

Error: write EPROTO 980B5FDB137F0000:error:0A00010B:SSL routines:ssl3_get_record:wrong version number:../deps/openssl/openssl/ssl/record/ssl3_record.c:354:

    at WriteWrap.onWriteComplete [as oncomplete] (node:internal/stream_base_commons:94:16) {

  errno: -71,

  code: 'EPROTO',

  syscall: 'write'

}

Anyway, your issue might be unrelated so I'll let you lads continue.

randomvalidOP•16mo ago

Same issue, bad gateway, but if I modify to 127.0.1.1 it works.

Meierschlumpf•16mo ago

Okay thanks for testing, so we are now sure it is actually happening because of changes in homarr i guess 😉

randomvalidOP•16mo ago

Welcome, if you need additional input from my side, just let me know! 🙂

Serenaphic•16mo ago

Found that the problem for the other issue is about the next dependency. The reason this doesn't work might be because of it too but for different reasons.

liquidguru•16mo ago

I access via a cloudflare tunnel, as I'm remote, and get the same error

Serenaphic•16mo ago

@randomvalid @salty 0.13.4 is out, would you be able to say if the problems are fixed now?

salty•16mo ago

Works fine my dev box now, I'll let you know if we get reports of it not working with Saltbox.

Manicraft1001•16mo ago

Awesome, thank you 👍

Serenaphic•16mo ago

Wonderful thank you for the feedback

randomvalidOP•16mo ago

Yes it's working, thank you for the quick fix! 🙂

Serenaphic•16mo ago

Thanks for confirming. Glad all is well

salty•16mo ago

Guessing this https://github.com/vercel/next.js/pull/54926 will fix it once it is merged for newer versions.

GitHub

Fix next-server not working properly when HOSTNAME is defined b...

What? Follow-up of #53131. This PR fixes issues caused by the previous PR, which are: req.url reporting the host's IP instead of the host itself, router.push sometimes uses the host's IP (c...

Serenaphic•16mo ago

Yeah that's also the one I saw when searching up on the subject

Thomas•15mo ago

This just got closed lmao, 4 hours ago

salty•15mo ago

hehe they did reference another PR that might fix it as well at least.

Gaming

Programming

502 Bad Gateway