Proxy crashes on startup with `invalid_dc_id` for negative DC IDs

[AliAlimohammadi]

Description

The proxy fails to start and enters an infinite crash/restart loop after Telegram began returning negative DC IDs (e.g. -203) in their configuration. The dc_to_pool_name/1 function in mtp_dc_pool.erl raises {invalid_dc_id, DcId} for any unrecognised DC ID, which propagates up through mtp_config and kills the application on startup.

Error

{invalid_dc_id,[{mtp_dc_pool,dc_to_pool_name,[-203],
  [{file,"/build/mtproto_proxy/src/mtp_dc_pool.erl"},{line,73}]}

Fix (confirmed working)

Instead of crashing on unknown DC IDs, skip them:

sed -i 's/error({invalid_dc_id, DcId})/begin io:format("Skipping invalid dc_id: ~p~n", [DcId]), skip end/' src/mtp_dc_pool.erl

Full steps to build and deploy the patched image:

git clone https://github.com/seriyps/mtproto_proxy.git
cd mtproto_proxy
sed -i 's/error({invalid_dc_id, DcId})/begin io:format("Skipping invalid dc_id: ~p~n", [DcId]), skip end/' src/mtp_dc_pool.erl
docker build -t mtproto-proxy-patched .
docker run -d \
  --name mtproto_proxy \
  --restart=always \
  --network host \
  mtproto-proxy-patched \
  -p <port> \
  -s <secret> \
  -t <tag> \
  -a dd

This fix makes the proxy skip any DC IDs it doesn't recognise rather than crashing, and the proxy starts and runs correctly.

Environment

• Image: seriyps/mtproto-proxy:latest
• Erlang: erts-10.3.5.19
• Release: 0.1.0
• Observed: February 2026 (following a Telegram infrastructure change)

[centuriononon]

Solved with #115

[AliAlimohammadi]

It's not fixed or maybe the Docker images are not updated on Docker Hub.
I used the patch I wrote above and then it works.

@centuriononon, @seriyps