Unable to run postgres:12.17 in gh actions

[indiVar0508]

hi

i was trying to use postgres:12.17 in gh actions, but not able to start service in my job i get below shown error while starting action job service container

I see that in local in docker the container restarts automatically but same is not happening for github service container.

        options: >-
          --health-cmd "pg_isready -U $$POSTGRES_USER -h $$(hostname -i)"
          --health-interval 60s
          --health-timeout 30s
          --health-retries 8

This is the Health command i am using.

[indiVar0508]

if i remove the healthcheck command to be only "pg_isready" it works, but then i get error message mentioned in #838

[tianon]

Instead of $$(hostname -i), can you try hard-coded 127.0.0.1?

[LaurentGoderre]

Doesn't omitting the host check itself?

[indiVar0508]

Yes thanks hardcoding 127.0.0.1 worked for me i also had to add -d db_name,
my final command looked like pg_isready -U user_name -d db_name -h 127.0.0.1 -p 5432
i was just wondering isn't this wrong that health marks container healthy but is silently failing in the background?
like if i just do pg_isready container is marked healthy but i get a warning log in DB that role doesn't exists like #838, even container should be marked unhealthy if command is wrong.

[tianon]

When you run this container the first time, it runs PostgreSQL twice -- once listening only locally in the container (on the appropriate unix socket) for running initdb and initialization scripts, then restarts listening externally. Thus, if you want to verify that the container is actually "ready" (fully initialized, all roles, etc), you have to connect via TCP.