gitlab-runner-helper tries to connect to tcp://docker:2376 instead of tcp://docker:2375

Summary

Not sure if it's a bug or just something silly I did.

I'm hitting a problem that When using docker executor with dind, gitlab-runner-help can't connect to the dind service. But the rest of the ci jobs run fine.

What is the current bug behavior?

gitlab-runner-helper throws below error at the beginning of job:

Error: Using Docker executor with image docker:stable ... Starting service docker:dind ... Pulling docker image docker:dind ... Using docker image sha256:e4157102c815f9ed8fec6118171866b13a331b6329ff8f85017f0bf7652917e9 for docker:dind ... Waiting for services to be up and running...

*** WARNING: Service runner-MyiRXyjg-project-841-concurrent-0-docker-0 probably didn't start properly.

Health check error: service "runner-MyiRXyjg-project-841-concurrent-0-docker-0-wait-for-service" timeout

Configs

\# ./gitlab-runner-helper --version Version: 12.2.0~beta.1736.gaa23dbcf Git revision: aa23dbcf Git branch: master GO version: go1.8.7 Built: 2019-07-18T11:18:38+0000 OS/Arch: linux/amd6

Some investigation I did

It turned out that gitlab-runner-helper was trying to connect to port 2376 but dind was only running on port 2375.

/ # gitlab-runner-helper health-check

waiting for TCP connection to 172.17.0.2:2376...

I understand that gitlab-runner-helper is getting the host and port from below environment variables, but is there a way to set it to use port 2375?

SERVICE_PORT_2376_TCP_ADDR=172.17.0.2

SERVICE_PORT_2376_TCP_PORT=2376

SERVICE_PORT_2376_TCP_PROTO=tcp

SERVICE_PORT_2376_TCP=tcp://172.17.0.2:2376

I set wait_for_services_timeout in runner config.toml to try to disable the check but it didn't work.

Cheers

Edited by Jessie Lin