Emby Client on Different Subnet from Server

[nominion 7]

I'm having an issue with my NVIDIA Shields logging in to the server.  The Sheilds are on one subnet (192.168.11.0/24) while the server is on another (192.168.12.0/24).  On the Shield, if you go back to the main menu and then go back to the Emby app, it can't find the pre-configured server.  It then presents a menu of users.  If I click on one of the user that I have setup, it logs in to the server right away.  

 

Has anyone else experienced this?

[ebr 16169]

Do you have the app configured to automatically log in to a specific user?

[nominion 7]

Yes, the app is configured to auto login.  I have it setup to stay logged in for up to 24 hours.  If I to back to the main Android menu on the Shield and go back in to Emby, that's when it breaks.  

 

It initially fails when trying to connect to the pre-configured server and presents the list of users.  If I click on the user, it logs in immediately, so I know the user/pass and the server IP are correct.

 

It's just weird that it doesn't work when launching the Emby app.  

[mwongjay 74]

I also had a similar experience and created a topic about the same issue. Below are the tests I did.

 

Below are excerpts from that topic describing the issue and what tests I tried

 

 

 

I have multiple networks. The server is on 192.168.20.0/24, the clients are all on 192.168.10.0/24. Whenever Emby is started from a fire tv client the loader icon begins spinning, freezes for 1-2 sec, then I receive the message "unable to connect to configured server <name>", and less than 2 seconds later it discovers and connects to the server. That process of starting up, loader freezing, error message, server discovery happens every time. 

 

 

Test cases:

 

When a user is set to auto log in: Log in, press Fire tv home button, open Emby, I am automatically logged in and works as expected.

When a user is set to auto log in: Log in, log out, open Emby, I receive issue stated earlier.

When user is NOT set to auto log in: The issue above is resolved. No trouble finding server and clicking a user works as expected.

 

It's odd this issue is only encountered when auto log in for a user is set and the user is required to authenticate. Do you know the reason this causes the issue? 

 

[nominion 7]

Thanks for adding that info to this thread.  Hopefully the developers are able to test this and find a resolution.  

[nominion 7]

Any updates?  Curious if anyone has been able to reproduce this issue as it continues for me and is very annoying.  Thanks!

[Luke 42077]

does the server have a static ip?

[nominion 7]

Yes, the server has a static IP address.  

[mwongjay 74]

Any updates?  Curious if anyone has been able to reproduce this issue as it continues for me and is very annoying.  Thanks!

 

For the record I still have this issue and while it is annoying there is really only a subset of users that really have this network setup. I anticipate this being an issue for a bit as it probably isn't a high priority which is unfortunate for us, but there are higher priority issues that affect many more users (picture issues, stream delays, etc). 

 

That being said my settings are to NOT automatically log in a user and logout after 30 min (or whatever that setting is in the fire tv app settings - unfortunately I'm not at home so I'm not sure exactly what it says). The result of these settings is when opening the Emby app after a significant period of inactivity it takes about 10 seconds to throw an error saying it can't find the server, then the users appear. If you close the app and reopen in a short period of time (couple hours) it has no problem finding the server and users are listed as expected.

[EarlyOnsetPooka 1]

I'm having this same issue 8 years later. Is there any setting for just increasing the wait time for the timeout message? Is that an emby client app issue or is it an android/os issue with shield, Google TV, and fire stick issue? Would it even be possible to just increase the allowed wait time in the app by 10 seconds or so?

[Luke 42077]

3 hours ago, EarlyOnsetPooka said:

I'm having this same issue 8 years later. Is there any setting for just increasing the wait time for the timeout message? Is that an emby client app issue or is it an android/os issue with shield, Google TV, and fire stick issue? Would it even be possible to just increase the allowed wait time in the app by 10 seconds or so?

Hi, how does this relate to multiple subnets? Can you please describe your issue in more detail? Thanks.

[EarlyOnsetPooka 1]

Absolutely. I'm also running multiple Emby client apps on their own lan in the 10.15.x.x subnet, while emby server is running in 10.20.x.x subnet. All rules in the unifi gateway are allowing communication, and the two DO communicate

Any client apps that run in 10.20.x.x connect immediately, while those on any other subnet will appear to timeout and immediately connect after the timeout message is displayed. Web clients don't have that issue, regardless of subject for me 

 

It's not affecting overall functionality so much as it's just an annoyance that the client apps on Google TV, Shield, or Fire stick will display the "unable to connect" toast notification just moments before the media list appears. There doesn't appear to be any lag after the clients connect. It's just the message appearing on the clients right before the media lists load on screen

[rbjtech 5284]

I run clients on multiple subnets/vlan's - I don't get the message.

How are you connecting on the clients ?   If via 'discovery' - then unless you are allowing udp /mdns relaying - then this will not cross vlan's.

If you use a direct IP or if you are running local dns, then a local dns/fqdn name on the client (ie http://1.2.3.4:8096 (or Port 8096) or http://myserver.lan:8096 (or Port 8096) then you do not need 'discovery' and it will instanly connect and not display the undiscovered message.  Port 8096) 

Edited June 12, 2025 by rbjtech

[EarlyOnsetPooka 1]

I'm using DNS on each client. It's emby.whatever.vlan:8096 and it still gives me the unable to connect message. Local DNS and VLAN rules are handled by Unifi gateway. Some clients are wireless and one is wired. It happens to all of them that aren't on the same vlan as emby 

[EarlyOnsetPooka 1]

1 hour ago, rbjtech said:

I run clients on multiple subnets/vlan's - I don't get the message.

How are you connecting on the clients ?   If via 'discovery' - then unless you are allowing udp /mdns relaying - then this will not cross vlan's.

If you use a direct IP or if you are running local dns, then a local dns/fqdn name on the client (ie http://1.2.3.4:8096 (or Port 8096) or http://myserver.lan:8096 (or Port 8096) then you do not need 'discovery' and it will instanly connect and not display the undiscovered message.  Port 8096) 

I absolutely forgot to mention that I'm auto connecting to last known user on all clients, as they're not located in shared spaces. I attempted to run it with logging in as a specific user, but no matter how many times I select the user and exit/reopen the emby client, it gives me a user selection screen instead of logging in as my chosen user. Discovery isn't even an option that I can find anywhere

Edited June 12, 2025 by EarlyOnsetPooka

[rbjtech 5284]

Hmm - so on the networking side - as long as the traffic is being allowed from the client (as the source) to the emby server (as the destination) on the correct port /protocol - then this should work fine.

Discovery will not work across vlans 'out the box' - you cannot just forward the udp requests, it doesn't work like that - but as you are not using discovery, it shouldn't make any difference.

I would probably start looking at the emby client - remove it, remove the cache, reboot the device and reload the client to start from scratch.  I'm assuming this is the AndroidTV client ?

[EarlyOnsetPooka 1]

16 hours ago, rbjtech said:

I run clients on multiple subnets/vlan's - I don't get the message.

How are you connecting on the clients ?   If via 'discovery' - then unless you are allowing udp /mdns relaying - then this will not cross vlan's.

If you use a direct IP or if you are running local dns, then a local dns/fqdn name on the client (ie http://1.2.3.4:8096 (or Port 8096) or http://myserver.lan:8096 (or Port 8096) then you do not need 'discovery' and it will instanly connect and not display the undiscovered message.  Port 8096) 

 

On 6/12/2025 at 12:27 AM, Luke said:

Hi, how does this relate to multiple subnets? Can you please describe your issue in more detail? Thanks.

 

6 minutes ago, rbjtech said:

Hmm - so on the networking side - as long as the traffic is being allowed from the client (as the source) to the emby server (as the destination) on the correct port /protocol - then this should work fine.

Discovery will not work across vlans 'out the box' - you cannot just forward the udp requests, it doesn't work like that - but as you are not using discovery, it shouldn't make any difference.

I would probably start looking at the emby client - remove it, remove the cache, reboot the device and reload the client to start from scratch.  I'm assuming this is the AndroidTV client ?

It is the android TV client for the Google TV box and the Shield, and probably the Amazon services as well. I've removed the server and reconnected to it on one of them, just because it's easy to access in the same room as my desktop PC that I use for management. It did occur to me that I've got the entire Emby server VLAN under a VPN and the clients are not. I'll have to change my VPN settings to exclude the Emby server VM and see if that makes a difference. Discovery is out anyway, as I'd rather use DNS/IP. I'll have to check into the VPN possibility later after work

[rbjtech 5284]

Have you checked DNS on those other VLAN's - do you have UDP 53 etc available as a local (and priority) lookup ?   Maybe it's trying to lookup your local emby fqdn on the internet or over the VPN rather than your gateway ?

 

[EarlyOnsetPooka 1]

1 minute ago, rbjtech said:

Have you checked DNS on those other VLAN's - do you have UDP 53 etc available as a local (and priority) lookup ?   Maybe it's trying to lookup your local emby fqdn on the internet or over the VPN rather than your gateway ?

 

I'll add that to the list of things to double check, but my emby server is under a completely different host name when exposed to family by remote

[EarlyOnsetPooka 1]

10 hours ago, rbjtech said:

Have you checked DNS on those other VLAN's - do you have UDP 53 etc available as a local (and priority) lookup ?   Maybe it's trying to lookup your local emby fqdn on the internet or over the VPN rather than your gateway ?

 

I've realized that I don't know HOW to check that and if you mean to check the emby server network or the iot client devices network.. All my VLANs are set up to use locally run pi-holes for DNS resolution, and I have a redirect in the pi-hole local DNS settings to forward the emby DNS name to the emby IP, just in case unifi forgets how it works 

[EarlyOnsetPooka 1]

Also, removing the emby server from the VPN made no difference, but the VPN definitely bypasses my local DNS resolver

[Luke 42077]

Quote

Also, removing the emby server from the VPN

Hi, what do you mean by this?

[Q-Droid 989]

Good back and forth in the thread but @EarlyOnsetPooka you haven't shared your Emby dashboard, network settings or any logs. More information that could help diagnose the problem.

 

[EarlyOnsetPooka 1]

13 hours ago, Luke said:

Hi, what do you mean by this?

I had the entire VLAN with my emby server running through a VPN, but changed it so the Emby server is not. The VPN devices don't use my local pihole DNS resolvers

Edited June 14, 2025 by EarlyOnsetPooka
Typo

[EarlyOnsetPooka 1]

2 hours ago, Q-Droid said:

Good back and forth in the thread but @EarlyOnsetPooka you haven't shared your Emby dashboard, network settings or any logs. More information that could help diagnose the problem.

 

Not sure how you mean to share my dashboard. Just a screenshot of the dashboard settings page? Is there a local file for network settings that I could share? THe only errors I'm seeing in my log files are filenames and directories for specific episodes not being resolved because of invalid characters.

I've got a log file with debug logging enabled from the past 15 minutes with two separate LAN client logins from the same device. Do I need to sanitize it at all and remove personal identifying info?