Problem
On Android, OpenNOW connects to GFN Bulgaria (NP-SOF-02) successfully, but the WebRTC media stream develops very high packet loss (typically ~40–80%). The official GeForce NOW client on the same network/device can use Bulgaria with normal/nearly zero loss.
Reproduction
- Open OpenNOW on Android.
- Connect without VPN and select/use Bulgaria / NP-SOF-02.
- Start a game.
- After the stream starts, packet loss rises to ~40–80% while ping may remain around 80–100 ms.
- OpenNOW shows the recovery sequence:
Recovering video → Reconnecting signaling → Waiting for offer → ICE_CHECKING.
- ICE then reaches
CONNECTED, but media can stall again; logs show media transport restart after a stall.
Relevant log observations
- ICE reaches
CONNECTED, so ICE negotiation itself succeeds.
- Media then stalls and OpenNOW triggers recovery/transport restart.
- One session showed ~65.6% loss and another showed very high loss with media stall/recovery.
- The affected endpoint was a Bulgaria/NP-SOF-02 session.
Environment
- Android device: KYV41
- OpenNOW: latest available version (reported by user as 0.7.5.119)
- Network: Zain Iraq small/local network, Basra, Iraq
- Official GFN on Bulgaria: works with nearly zero packet loss
- OpenNOW codecs/settings tested: H.264/H.265 and L4S; issue persists
- Restarting the network did not resolve it
Suspected area
The issue appears to occur after ICE connects, in the media transport path. Please investigate MediaConnectionInfo / ICE candidate endpoint rewriting and WebRTC media transport/recovery for this type of network. I can provide the full OpenNOW log if needed.
Please let me know if additional logs or diagnostics are required.
Problem
On Android, OpenNOW connects to GFN Bulgaria (NP-SOF-02) successfully, but the WebRTC media stream develops very high packet loss (typically ~40–80%). The official GeForce NOW client on the same network/device can use Bulgaria with normal/nearly zero loss.
Reproduction
Recovering video→Reconnecting signaling→Waiting for offer→ICE_CHECKING.CONNECTED, but media can stall again; logs show media transport restart after a stall.Relevant log observations
CONNECTED, so ICE negotiation itself succeeds.Environment
Suspected area
The issue appears to occur after ICE connects, in the media transport path. Please investigate MediaConnectionInfo / ICE candidate endpoint rewriting and WebRTC media transport/recovery for this type of network. I can provide the full OpenNOW log if needed.
Please let me know if additional logs or diagnostics are required.