Last days i was working with the geonode-project setup a lot.
I found few inconveniences which makes debugging the stack very troublesome:
1) entrypoint.sh redirects logs
in every place, entrypoint invokes some action the output is redirected to the file. i.e:
/usr/local/bin/invoke migrations >> /usr/src/ibf_meteoblue_host/invoke.log
If the exception kills the container it is falling in the endless restart loop which makes reading the logs very unpleasant.
The logs should always be redirected to STDOUT, the file can be second logs storage. if something is wrong simply docker logs is enough to debug issue moreover STDOUT is easy to process from outside using any log harvester
Last days i was working with the geonode-project setup a lot. I found few inconveniences which makes debugging the stack very troublesome: 1) entrypoint.sh redirects logs
https://github.com/GeoNode/geonode-project/blob/master/entrypoint.sh
in every place, entrypoint invokes some action the output is redirected to the file. i.e: /usr/local/bin/invoke migrations >> /usr/src/ibf_meteoblue_host/invoke.log If the exception kills the container it is falling in the endless restart loop which makes reading the logs very unpleasant.
The logs should always be redirected to STDOUT, the file can be second logs storage. if something is wrong simply docker logs is enough to debug issue moreover STDOUT is easy to process from outside using any log harvester