Closed solareon closed 2 weeks ago
Yes, the stop signal is not handled correctly now.
In launch.sh
the trap kill_corekeeperserver EXIT
should run when the container receives the stop signal. Unfortunately, the trap
is not called when the container is stopped via docker stop
or docker compose down
.
This is still an open problem as part of issue #22 \ I had no time yet to further investigate this problem.
So adding tini to the container/entrypoint fixes all of that. This PR adds that. When you stop the container the container stops instantly as the sigterm is passed to the forked process at https://github.com/escapingnetwork/core-keeper-dedicated/blob/45c47d2e971c2320dc4e039250cb15cd9ef83800/launch.sh#L72.
Adds tini to support passing through application signals allowing server to stop without waiting for timeout.