What type of hardware are you using: Old laptop or computer What YunoHost version are you running: 12.1.39 Are you in a special context or did you perform specific tweaking on your YunoHost instance ?: I change my internet provider
Describe your issue
I am following the discussion started at [Synapse] Instant messaging server matrix network - #169 by croulibri
I changed my Internet provider from SFR to Free and so I got a new “internet box” (modem/router).
I set up port forwarding as I always do when changing internet providers (and this isn’t the first time…).
My Yunohost server is accessible via http(s) and for Matrix messages (Nextcloud, Dokuwiki, Piwigo, and Synapse are working and accessible outside home).
So ports 443, 80, 908 (for SSH), 5353, and even 5349 are open.
But ports 5350 and 49153-49193 are not open. And I need them for Element call.
Both the Yunohost firewall and my Box redirects are in place (for instance 5349 is configured the same as 5350, but 5349 is open and 5350 closed!). I can see that the ports are closed by testing with
nc -vvv -w 10 -z my.server.fr 5349
Connection to my.server.fr (XX.EE.DD.EEE.EE) 5349 port [tcp/*] succeeded!
nc -vvv -w 10 -z my.server.fr 5350
nc: connect to my.server.fr (XX.EE.DD.EEE.EE) port 5350 (tcp) failed: Connection refused
nc: connect to my.server.fr (ipv6 adress) port 5350 (tcp) failed: Network is unreachable
I can’t understand why it’s not working.
I tried deleting the entries in the firewall, then putting them back. I deleted the Internet Box redirects, then put them back. I restarted the box and server each time.
Nothing works. Port are closed. Element calls doesn’t work.
@Josue helped me trying to look at the log and console when calling through web browser.
I have set a DMZ forwarding all ports to my server. I have checked on several ports and the DMZ seems effective.
This narrow down the issue to the firewall or conf of the Yunohost / Synapse server…
I have changed the IP indicated in the Coturn conf as it is writtent inside the settings. With new internet provider comes new fix IP.
There is an issue that make the Element call not established, even if it rings on both sides.
If a matrix.org account calls me, it works. But if an account on my server call another account on my server or a matrix.org account, it doesn’t work…
Could it be indeed not an issue of Port forwarding / Port opening but rather an issue of TURN server?
Bonjour
The command `sudo ss -tulpn | grep 5350` doesn't give me anything in return. The same with `sudo ss -tulpn | grep 5349`
I have a full stack ipv4, and Yunohost application using other port are working (Synapse on 8448, Nextckoud or Dokuwiki on 443,…). So redirection of Port seems to work.
I was wondering about the change I should do when I change IP. I understand I have to change settings of Coturn, but what should I do exactly and what other changes should I do?
So coturn isn’t listening ? What does systemctl status coturn return ?
There’s a Coturn_config_rotate.sh in /home/yunohost.app/coturn. But maybe yunohost app upgrade coturn --force could do the trick to regen coturn’s configuration.
Thanks for your advices!
``
sudo systemctl status coturn × coturn.service - coTURN STUN/TURN Server
Loaded: loaded (/lib/systemd/system/coturn.service; enabled; preset: enabled)
Active: failed (Result: exit-code) since Sat 2026-02-14 00:03:27 CET; 13h ago
Docs: man:coturn(1)
man:turnadmin(1)
man:turnserver(1)
Process: 976 ExecStart=/usr/bin/turnserver -c /etc/turnserver.conf --pidfile= (code=exited, status=255/EXCEPTIO>
Main PID: 976 (code=exited, status=255/EXCEPTION)
CPU: 17ms
Feb 14 00:03:27 my.server.com systemd[1]: coturn.service: Main process exited, code=exited, status=255/EXCEPTION
Feb 14 00:03:27 my.server.com systemd[1]: coturn.service: Failed with result ‘exit-code’.
Feb 14 00:03:27 my.server.com systemd[1]: Failed to start coturn.service - coTURN STUN/TURN Server.
Feb 14 00:03:27 my.server.com systemd[1]: coturn.service: Scheduled restart job, restart counter is at 5.
Feb 14 00:03:27 my.server.com systemd[1]: Stopped coturn.service - coTURN STUN/TURN Server.
Feb 14 00:03:27 my.server.com systemd[1]: coturn.service: Start request repeated too quickly.
Feb 14 00:03:27 my.server.com systemd[1]: coturn.service: Failed with result ‘exit-code’.
Feb 14 00:03:27 my.server.com systemd[1]: Failed to start coturn.service - coTURN STUN/TURN Server.
lines 1-18/18 (END)
``
I restarted it and I get
``
sudo systemctl status coturn ● coturn.service - coTURN STUN/TURN Server
Loaded: loaded (/lib/systemd/system/coturn.service; enabled; preset: enabled)
Active: active (running) since Sat 2026-02-14 13:12:26 CET; 1min 37s ago
Docs: man:coturn(1)
man:turnadmin(1)
man:turnserver(1)
Main PID: 21174 (turnserver)
Tasks: 9 (limit: 9213)
Memory: 7.0M
CPU: 142ms
CGroup: /system.slice/coturn.service
└─21174 /usr/bin/turnserver -c /etc/turnserver.conf --pidfile=
Feb 14 13:12:26 my.server.com turnserver[21174]: 0: : turn server id=2 created
Feb 14 13:12:26 my.server.com turnserver[21174]: 0: : turn server id=0 created
Feb 14 13:12:26 my.server.com turnserver[21174]: 0: : IO method (general relay thread): epoll (with changelist)
Feb 14 13:12:26 my.server.com turnserver[21174]: 0: : turn server id=3 created
Feb 14 13:12:26 my.server.com turnserver[21174]: 0: : Total General servers: 4
Feb 14 13:12:26 my.server.com turnserver[21174]: 0: : IO method (auth thread): epoll (with changelist)
Feb 14 13:12:26 my.server.com turnserver[21174]: 0: : IO method (auth thread): epoll (with changelist)
Feb 14 13:12:26 my.server.com turnserver[21174]: 0: : IO method (admin thread): epoll (with changelist)
Feb 14 13:12:26 my.server.com turnserver[21174]: 0: : SQLite DB connection success: /var/lib/turn/turndb
Feb 14 13:12:26 my.server.com systemd[1]: Started coturn.service - coTURN STUN/TURN Server.
``
I don’t have a Coturn_config_rotate.sh in /home/yunohost.app/coturn because coturn is installed with/by Synapse in my case, following the setting of @Josue
So I don’t know exactly if I can/should force a regen of coturn’s configuration.
with all exposed should be opened, so 8448 TCP, 5349 BOTH, 5350 BOTH, 7881 TCP
And for all these PORT it seems OK
5350 doesn’t answers, but I understoofd it might be because nothing listning on this port.
And for the port port 49153-49193, Josue told me it is more difficult to assess as Livekit or Coturn assign an IP in this range to client when needed. So all this range of port if not always used…
I might be mistaken, but that’s why I am looking on the Livekit and coturn config update when a change of IP happen. Is there something to do indeed?
Another element… when I try to call through Element Call, my correspondent ring.
The correspondent answer the call, call goes in full screen with “waiting for media” written in the middle. After few seconds, the call abort.
So it seems tha both correspondant found themselves, but the media stream can not be shared.
This is a very subjective and “not expert at all” analysis, my feeling.
What could prevent media stream to be shared. It is Livekit (Element Call) and Coturn (Element Classic), or another system?
On which port these media stream are exchanged?
I update Synapse from 1.144.0~ynh1 to 1.147.1~ynh1 as suggested with the command sudo yunohost app upgradeand now Element call are working again.
So I guess the problem was config problem, but I don’t know where the problem was.
It is solved and everything works again.
Thanks @otm33 and also @Josue for your timely and very informative support!
I think I have gained some knowledge on port management (only some as this seems quite technical).
