Odd Issue On Select Devices without Docker ports published

[CassTG 98]

Okay First ill explain the simple setup

• Emby latest setup in Docker on a Remote Cloud infrastructure
• All containers are sat behind Linuxserver Swag Proxy using ZeroSSL certs
• Swag port published is 443, all domains for containers go via HTTPS
• Swag passes traffic up to Emby using standard proxy-conf
• Emby setup for public port 443
• Dashboard shows correct published domain and ports for Emby Connect
• All devices discussed here are signed in via Emby Connect
• Preroll files via vantage point are local, my media is via rclone mount
• Issue happens without VantagePoint installed
• One affected device is GTKING Pro - running Android 9 and app 3.2.05 (will update when next round)
• When i mention PUBLISHED PORTS i am reffering to the setup of the cokcer container itself, not the Emby network settings

 

The Issue

A select couple of devices will just not play any media if the emby docker is setup without ports 8096 and 8920 published. As far as i understand as everything is going through the Swag proxy via port 443, so this should not matter

The odd thing is this, all my devices at home which included Nvidia Shield, Firecube, Galaxy Ultra s21, Samsung and LG tv Apps all work as expected when the ports are not published.

A friend who access via Ipad, and LG Tv does not have an issue

Now a possible red herring here - we wre both on BT ISP

But My Mum and another Friend who (red herring?) are on different ISP cannot actually play any main feature.

I was working through this last night and remote controlled their system to see what was happening. I had faced this issue before so i thought before rebuilding the container i would see it for myself.

They can access the server via the app, browse etc no issues, the Emby logs and Swag proxy logs show this interaction okay. So i started to play the film via remote and i noticed the 2 pre rolls worked perfectly but when it got to the end of the 2nd preroll, it showed the film starting and then just stuck there, not progressing just paused on 1st second.

 

The Logs pt 1

I checked the logs and nothing of any interest showed up, it was just updating playing process as normal just nothing was actually playing.

 

Rebuilt Docker

So having seen this before, i decided to nuke the container and re publish the 8920 and 8096 ports in the rebuild. Obviously the SSL port wouldnt work as thats set to 443 in the network section so this is correct.

I remote controlled in and restarted the same film, the 2 pre rolls worked as expected and then the film loaded instantly and proceeded to play normally.

For info at no point did any transcode take place all was direct Playing.

 

The Logs pt 2

So i checked both Nginx access logs and Emby Logs, and again no issues visible, whats interesting is the the Swag proxy logs were still reporting all player status updates etc via itself, so the device is still connecting correctly via embyconnect published domain and ports.

 

TLDR

Linuxserver Emby docker is behind Linuxserver Swag proxy, all port and network settings are correct, this only affects 1 or 2 devices where playback of main feature will not start (just appears to look like it's buffering) unless the docker ports are published which shouldn't be required as all traffic is going through port 443 on the Swag Nginx proxy container. Even when ports are published, all traffic still routes via the Swag Proxy.

I kinda don't like publishing ports on the server when i don't have to hence why all my containers have no public ports published so that everything goes through the Swag proxy

I just can't get my head around why just 1 or 2 devices will appear to endlessly buffer if the ports are not published at all, and the Logs show no errors at all when the media pauses at the start just the standard player status updates nothing else.

 

 

 

 

Edited October 10, 2021 by CassTG

[Luke 37053]

HI, did you try doing a manual connection on these devices to manually enter the server connection information? What devices, OS versions and Emby apps are being used?

[CassTG 98]

Both people who have the hit and miss issue connect via Emby Connect.

Devices affected are only GT-King Pro (pls note nothing pro about it, overpriced junk) running on some shoddy Beelink Firmware and a new Samsung TV (although their firestick had the same issue which is whats odd)

However I have Zero issues on Shield, Firecube, S21 Ultra, Samsung and LG Tv's

To be honest normally i look for the logic to solve the issue but there is no logic here, all traffic still goes through the proxy fine, even if the ports are published, thats why i cant get my head round why the main feature wont start if they are not published.

Anyway's thinking about it, it doesn't really matter if they are published as if you know the domain / ip you can find open ports anyway thus reaching the login page, will just add a fail2ban entry on it (the proxy is already protected)

[Luke 37053]

It would still be helpful to ask them to try a manual connection the next time they run into this.

[CassTG 98]

On 15/10/2021 at 16:30, Luke said:

It would still be helpful to ask them to try a manual connection the next time they run into this.

I am popping round to one next Friday so i will log it in manually and then remove the ports to see if the issue occurs

[Luke 37053]

Let us know how you get on. Thanks.

[CassTG 98]

Okay

The box was setup using Beelinks Original latest firmware. So standard android app not ATV. so login is via username and password not emby.media/pin. buffering was bad, updated to latest beta app and that was as bad.

So i brought the box back (gave them a firecube instead which worked like a treat) and wiped the OS putting on a custom ATV ROM, so far it streams like a champ.

So in this instance i think it goes back to my original thoughts on Beelink, Good looking hardware running half baked generic OS thats lucky it even boots.

[Luke 37053]

Thanks for the feedback.