@jvanulde noticed that the pygeoapi container stops trying to spawn after n number of hours when the add_data.sh takes too long to run.
Need to figure out:
Is there is a docker-compose default limit on the number of times a docker container is allowed to restart from error? See in Compose file version 3 reference:
~Or any such similar in the pygeoapi Docker image?~
Consider adding a quick test, or use a tool such as wait-for-it (available in the Debian sid based pygeoapi Docker image), rather than running pygeoapi generate-openapi-document -c ${PYGEOAPI_CONFIG} > ${PYGEOAPI_OPENAPI} full-on in /entrypoint.sh:
See also:
73
107
@jvanulde noticed that the pygeoapi container stops trying to spawn after n number of hours when the add_data.sh takes too long to run.
Need to figure out:
restart
(from version 2 but still valid in version 3): https://docs.docker.com/compose/compose-file/compose-file-v3/#restartrestart_policy
underdeploy
: https://docs.docker.com/compose/compose-file/compose-file-v3/#restart_policyConsider adding a quick test, or use a tool such as
wait-for-it
(available in the Debian sid based pygeoapi Docker image), rather than runningpygeoapi generate-openapi-document -c ${PYGEOAPI_CONFIG} > ${PYGEOAPI_OPENAPI}
full-on in /entrypoint.sh:from https://github.com/geopython/pygeoapi/blob/master/docker/entrypoint.sh#L62-L63 at the time of writing.
This will hopefully makes pygeoapi container restarts more gracefully and less verbose, unlike the current harmless but scary-looking messages below: