Fellow Gentoo user! Kudos.

Well, here is the relevant part then, sorry if it was not clear:

• Jellyfin will not play well with reverse proxy auth. While the web interface can be put behind it, the API endpoints will need to be excluded from the authentication (IIRC there are some examples on the web) but the web part will stil force you to double login and canot identify the proxy auth passed down to it.
• Jellyfin do support OIDC providers such Authelia and it’s perfectly possible to link the two, in this case as i was pointing out, Jellyfin will still use it’s own authentication login window and user management, so the proxy does not need to be modified.

TLDR: proxy auth doesnt work with Jellyfin, OIDC yes and it bypassess proxy, so in both cases proxy will not be involved.

This is my jellyfin nginx setup: https://wiki.gardiol.org/doku.php?id=services:jellyfin#reverse-proxy_configuration

currently i don’t use any proxy related authentication because i need to find the time to work with the plugins in Jellyfin. I don’t have any chromecast, but i do regularly use the Android Jellyfin app just fine.

I expect, using the OIDC plugin in jellyfin, that Jellyfin will still manage the login via Authelia itself, so i do not expect much changes in NGINX config (except, maybe, adding the endpoints).

Never found a service that don’t work with nginx reverse proxy.

My jelly fin does.

Don’t run photoprims tough…

You might use LDAP, but its total overkill.

I have not yet worked jellyfin with authelia, but its more or less the last piece and I don’t really care so far if its left out.

A good reverse proxy with https is mandatory, so start with that one. I mean, from all point of views, not login.

I have all my services behing nginx, then authelia linked to nginx. Some stuff works only with basic auth. Most works with headers anyway, so natively with authelia. Some bitches don’t, so I disable authelia for them. Annoying, but I have only four users so there is not much to keep in sync.

They actually do, i am down the same path recently and installing authelia was the best choice I made. Still working on it.

But most stvies support either basic auth, headers auth, oidc or similar approaches. Very few don’t.

Ok, I have a web browser on a locked down device and nothing else: how do I print a pdf or a photo using IPP?

I have: a camera, a browser, a file manager (kind of, think of an iPhone or some stock android business device) and I need to print a photo taken with the camera or a pdf file sent to me via email or WhatsApp?

The device is connected to the WiFi guest network with limited internet access (if any) and as only available service a server with port 443 open (a reverse proxy on that, captive portal and such).

In my experience, there is no way to print via cups in this configuration. Maybe I am wrong?

It still requires the device to be capable to print…

And the user to find the printer select it and so on. And must expose more ports on the network beside 443…

So, indeed cups is a great solution, but not to the problem I want to solve.

I do use cups in fact for the trusted part of the network, driverless printing for windows and Linux. Android doesn’t even need cups since it picks up the printer directly from the printer itself (AirPrint or whatevee that’s called).

I known cups can share printers and queues.

What is unclear?

I don’t want to pull drivers or install cups on devices. I want to print from anywhere just uploading a file to a web page.

If I have lots of devices or just want to let somebody print from his phone/tablet without installing or configuring anything…

With cups I still need to touch the system or the device somehow to let it print.

Yes, this is what I am afraid of… There is nothing out there for this task.

Hope to find something or maybe try to create something using lpr on the background… But this is the las hope as I have little time.

I want to print from a web page: upload the file, hit print button.

In this way I can print from whatever device I want even without any driver installed or configuration.

Yes, sure I am… Would probably prefer a bash CGI because I like challenges :)

(Author and such would be managed by my reverse proxy)

But I would prefer something already baked if it exist

You get a real IP? Its been cg-nat with every provider for the last many many years in Italy.

I got a cheap vps and just run some reverse tunnels to map ports from it to my home server going trough my cg-nat.

Don’t. Backup data, reinstall software.

Answer.

No, because i did it over long time since I encoded only during the day hours and a few episodes at a time.

I can say that I fit a good 30% more episodes in the same space, but at the same time I also have added movies and reduces sizes too, so hard to tell reasonably.

I can say that all my collection was mostly h264 before.

Maybe I have a different point of view here, but I have actively converted all my TV series to AV1 and will probably to the same to most of my movies.

The space saving is huge, and the quality is identical to my eyes and hardware. True that storage is cheaper than ever, but this is not a reason valid to waste it anyway.

I have only 6TB of storage for my media and the power needed to run additional disks would only be waste on the long run, and so buying new bogger disks would be a waste for stuff I don’t wantch often (more like hoarding than…).

So AV1 is the way. Software encoding is the best quality, I have heard, rather than hardware encoding. As for playback, I have a fire stick with AV1 support that works flawlessly, so.

Edit: I have FV at home, so converting to AV1 during daylight is actually free for me.

Italian content? Some, but not much. Way less that torrents.

I speak from experience.

For anybody waiting for invites: i have up and just waited until Black Friday deals and openings: you need to head to “the other site” and watch the deals and opening times and you can get anything you want.

Also, invites are not free: you still need to pay the subscription!

They get DMCA’d regularly and content get removed on Usenet as well. But the fact that they have to report literally thousands of individual files every time make it slow and inefficient. People will just reload the same item many times and it’s always there.

Each copy, each single file in which the copy is split needs to be identified and asked for removal. Compared to torrents, it’s a long and complex task.

Well, my experience is that unless you set static IP+DNS in android WiFi advanced networks, it will not obey the dhcp option 6.

LineageOS, vanilla with mind the gapps

Yes, perfectly… My guess is android bypass local resolver and goes via DoH l, which sucks hard

Old thread, but I have somehow solved by reinstalling unbound and nuking the old config file…

Thanks! This explains a few things… But not why Android is IGNORING my DNS pushed via DHCP even if private DNS is disabled…

It’s blatantly ignoring the DNS i set via DHCP it seems. Only if i set it manually (static) it will use it! I have no subnets

Private DNS is disabled.

And even ping from tmux fails…

Install some DNS test app for Android and check that it does get resolved.

My will resolve the home server address to 0.0.0.0, and I get the same network error.

I have a very similar issue. Seems like Android will bypass your DNS resolver and thus cannot resolve your local names.

I have my home services on “home.my domain.com” accessible from outside and re-mapped to “192.168.0.1” (my internal server IP) at home, and all PCs can access it while all android phones can only resolve to the public IP.

I feel it’s something related to DoT or similar but haven’t yet dig in that.

Been using USB since forever. I have one USB3 jbod with four SSDs (2 RAID1 + 2 RAID1) and an USB-C double enclosure with two spinning disks (RAID1 as well)

Setup is running flawless since 2016 at the very least, but I was using USB even before that timeframe.

All the RAIDs are Linux software RAID.

As codecs go, I love AV1. More and more hardware support everywhere, it seems more future proof than h265 even at this point.

Rip at native DVD resolution and don’t forget ALL the audio tracks, as many English users always forget about the rest of the world.

My guess you had broken cables or defective connectors. Because even on cat5 (not cat5e) you should get much more than 7mbit, or did you have coaxial? LoL.

In my experience 90% are plugs, specially if you crimped yourself with Chinese tools

Had the same issues, it was heat.

Cool down your server, add a fan or a cooler…

I added a usb-powered fan sucking cooler air from outside the server area directly blowing it on the chassis.

That fixed for me.

Indeed you are, but if you release it you probably want people to use it, otherwise why release at all? Going github/gitlab way just make that easier.

You are free to do what you want ofc, just my thoughts.

Why not publish on github or gitlab? Or setup a gitea/similar website and post it?

I did so with a few of my small stuff.

I have read that kopia has corrupted Dara systematically in the past. What’s your experience with it?

Restic has a neat web GUI called Backrest

Thats perfect

Restic or Borg on your side, a safe and remote destination on the other side.

use restic, with backrest web GUI, and cannot be happier.

As for remote site, I use a remote machine I rent, but there are plenty of providers around, shop a bit… Or find a friend for reciprocal backup?

Synchting is the way… Been using it since ever. Easy to setup, works flawlessly, doesn’t get into the way.

I use it for server, PC, laptop and android devices.

… Then you want to backup (Borg, restic…) your synched files of course …

What?

Nginx “just works™” had never got into the way, its been rock solid and has not changed significantly over the years.

Why would I need something else?