container that is running postgrest server cannot be stopped with SIG_INT (ctrl+c)

Start hivemind server (change database url according to Your setup)

docker run -d --network haf_test  --name=hivemind-postgrest-server registry.gitlab.syncad.com/hive/hivemind:878c8168 postgrest-server --database-url=postgresql://hivemind@172.17.0.2:5432/haf_block_log

then stop it and check return value:

docker stop hivemind-postgrest-server
docker inspect  hivemind-postgrest-server --format='{{.State.ExitCode}}'

It returns 137 because docker stops container with SIG_KILL when could not end it with SIG_INT.

Consequences: postgrest is not cleanly closed, affects HAF APi Node setup by adding delay for tries with SIG_INT during compose is being down

Added issue to the postgrest project: https://github.com/PostgREST/postgrest/issues/3884

assigned to @asuch

Although @asuch found a way to break postgrest with SIG_INT in our docker setup there is still an issue with the returned value, postgrest broken with SIG_INT returns correct exit code 130 (128+2(SIG_INT), but when docker starts it it returns 255:

docker build -t hivemind_local .
docker run --rm -it  -e PGRST_DB_URI=postgresql://hivemind@172.17.0.2:5432/haf_block_log -e PGRST_LOG_LEVEL=debug --name=local_test --entrypoint postgrest hivemind_local

press'Ctrl+C' and then:

$ echo $?
$ 255

The solution is to use tini (—-init docker option), but this suggests that Postgrest may have some problems passing signals to its threads/child processes, and docker cannot be closed gracefully.

Commit which adds exec to start postgrest script: 4740d4e7

When I run directly postgrest I have 130 exit code:

[someone@udooboltv8 hivemind]$ postgrest
28/Jan/2025:13:36:12 +0100: Attempting to connect to the database...
28/Jan/2025:13:36:12 +0100: Connection successful
28/Jan/2025:13:36:12 +0100: Admin server listening on port 3001
28/Jan/2025:13:36:12 +0100: Listening on port 8080
28/Jan/2025:13:36:12 +0100: Config reloaded
28/Jan/2025:13:36:12 +0100: Listening for notifications on the pgrst channel
28/Jan/2025:13:36:12 +0100: Schema cache loaded
^C
[someone@udooboltv8 hivemind]$ echo $?
130

When I try to run postgrest via docker, I need to use --init, otherwise I will receive 255 exit code.

[someone@udooboltv8 hivemind]$ docker run --init -d --name=hivemind-postgrest-server registry.gitlab.syncad.com/hive/hivemind:local-hivemind-develop server --database-url=postgresql://hivemind@172.17.0.2:5432/haf_block_log
fb1d66b12287ddcd89abc4f9f5a3e20ce9e398d7431cdb15d1df464cc5f14d8c
[someone@udooboltv8 hivemind]$ docker stop hivemind-postgrest-server
hivemind-postgrest-server
[someone@udooboltv8 hivemind]$ docker inspect  hivemind-postgrest-server --format='{{.State.ExitCode}}'
130
[someone@udooboltv8 hivemind]$ docker container rm fb1d66b12287ddcd89abc4f9f5a3e20ce9e398d7431cdb15d1df464cc5f14d8c
fb1d66b12287ddcd89abc4f9f5a3e20ce9e398d7431cdb15d1df464cc5f14d8c
[someone@udooboltv8 hivemind]$ docker run -d --name=hivemind-postgrest-server registry.gitlab.syncad.com/hive/hivemind:local-hivemind-develop server --database-url=postgresql://hivemind@172.17.0.2:5432/haf_block_log
630673f3206649b829c90ff96cb7b8dad169ba9cae56e39657064721bb078be3
[someone@udooboltv8 hivemind]$ docker stop hivemind-postgrest-server
hivemind-postgrest-server
[someone@udooboltv8 hivemind]$ docker inspect  hivemind-postgrest-server --format='{{.State.ExitCode}}'
255

I added an issue to postgrest Github: https://github.com/PostgREST/postgrest/issues/3884

changed the description