You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am setting up a LiveKit and SIP servers on remote machine and having a connection problem. The latest source codes were fetched from GitHub, compiled with mage.
In a local environment (Linux, Ubuntu), it works fine.
Now, I want to deploy it to a Cloud environment (Compute Engine, GCP). All TCP/UDP ports are open on GCP Firewall.
Here are the configurations I used:
When I connect via a SIP client (Zoiper) using TCP protocol, everything works correctly, and I can establish voice communication.
However, when I switch to a more applicable SIP transport protocol, i.e., UDP, I am unable to establish a successful connection.
When I checked the logs on the client side, I see the SIP INVITE (outgoing message), then receiving 100 and 180 messages from the remote machine. However, the 200 OK message is not arriving at my SIP client - thus, the connection cannot be established.
I checked the logs on the Compute Engine side, and it appears that the 100 and 180 messages are sent to the correct public IP of my router, while the 200 message is being sent to my private IP address (inside NAT): 10.132.0.18.5060 > 192.168.0.22.64230. Please see the logs below.
I also tried a SIP client using a public IP, still the 200 message seems to be sent to the remote machine's local IP address range (10.132.0.18). When TCP protocol is used, then the destination IP address for 200 OK message is correct (public IP of my router).
I also tried using LiveKit's production deployment configuration (using Caddy, +STUN/TURN servers) - but the issue persists.
When I set up a simple SIP server on the remote machine (no STUN, TURN, IP discovery) based on emiago/sipgo, )to test firewalls, ports etc.), I am able to establish a successful connection with UDP transport (regardless if the Client is public or within NAT).
I'm not sure if this issue is related to the SIP server configuration or the LiveKit server RTP configuration or some internal bug. I've tried many configurations, but haven't had any success so far. I do not see any problems/errors in SIP server logs, while livekit server logs are quite hard to analyse.
Could you please take a look and possibly provide advice or fix the issue if it exists?
If helpful, I can also provide SIP logs and livekit logs.
I've build the executables directly from the latest codes pulled yesterday (21.10.2024) from github (livekit/livekit and livekit/sip) - to have the latest codes. But also experiencing same issue with older codes.
I am setting up a LiveKit and SIP servers on remote machine and having a connection problem. The latest source codes were fetched from GitHub, compiled with mage.
In a local environment (Linux, Ubuntu), it works fine.
Now, I want to deploy it to a Cloud environment (Compute Engine, GCP). All TCP/UDP ports are open on GCP Firewall.
Here are the configurations I used:
sip_config.yaml
livekit.yaml
I am running the servers via command lines:
When I connect via a SIP client (Zoiper) using TCP protocol, everything works correctly, and I can establish voice communication.
However, when I switch to a more applicable SIP transport protocol, i.e., UDP, I am unable to establish a successful connection.
When I checked the logs on the client side, I see the SIP INVITE (outgoing message), then receiving 100 and 180 messages from the remote machine. However, the 200 OK message is not arriving at my SIP client - thus, the connection cannot be established.
I checked the logs on the Compute Engine side, and it appears that the 100 and 180 messages are sent to the correct public IP of my router, while the 200 message is being sent to my private IP address (inside NAT): 10.132.0.18.5060 > 192.168.0.22.64230. Please see the logs below.
I also tried a SIP client using a public IP, still the 200 message seems to be sent to the remote machine's local IP address range (10.132.0.18). When TCP protocol is used, then the destination IP address for 200 OK message is correct (public IP of my router).
I also tried using LiveKit's production deployment configuration (using Caddy, +STUN/TURN servers) - but the issue persists.
When I set up a simple SIP server on the remote machine (no STUN, TURN, IP discovery) based on emiago/sipgo, )to test firewalls, ports etc.), I am able to establish a successful connection with UDP transport (regardless if the Client is public or within NAT).
I'm not sure if this issue is related to the SIP server configuration or the LiveKit server RTP configuration or some internal bug. I've tried many configurations, but haven't had any success so far. I do not see any problems/errors in SIP server logs, while livekit server logs are quite hard to analyse.
Could you please take a look and possibly provide advice or fix the issue if it exists?
If helpful, I can also provide SIP logs and livekit logs.
Additional info
ifconfig:
Logs from Remote Compute Engine:
The text was updated successfully, but these errors were encountered: