previously the log (only the exceptions) were log
into /var/log/uwsgi/uwsgi.log
this is disturbing for the admins:
* they see an internal error on HTTP port
* no log where they are expected (docker logs)
this commit fixes this issue
In commit 5a7b12e we normalized settings.yml and c6a5cc019 and de5a8ee7d we
changed instance name to ``instance_name: "SearXNG"``.
This patch adjust the sed expressions to modify:
dockerfiles/docker-entrypoint.sh
Closes: https://github.com/searxng/searxng/issues/876
Suggested-by: @neiaberau
Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
Add script docker-entrypoint.sh to shellcheck and try to _simplify_ and
_normalize_ some parts:
- fix issues reported by shellcheck
- don't mix tab and space indent
- command 'help' replaced by '-h': ./dockerfiles/docker-entrypoint.sh -h
- replace printf in help() by 'cat <<EOF'
Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
See #1561 , use uwsgi and Alpine Linux
Volume:
/var/log/uwsgi contains error log for 2 days (file uwsgi.log)
/etc/searx contains the settings.yml and uwsgi.ini files.
The docker image creates them if they don't exist.
The two files can be modified after the first run. See below.
Environement variables:
MORTY_URL : external URL of Morty
MORTY_KEY : base64 encoded key
BASE_URL : external URL of Searx
BIND_ADDRESS : internal HTTP port to listen to
Labels : org.label-schema.schema.*
Parameters:
-h : display this help
-d : will update the settings and quit immediately (settings.yml and uwsgi.ini)
-f : always update the settings (previous version saved with suffix .old).
without this parameter, the new settings are copied with suffix .new
When the Docker image contains newer settings:
- without -f parameter: the new versions are copied to /etc/searx/settings.yml.new and /etc/searx/uwsgi.ini.new.
- with -f parameter: the old versions are renamed with .old suffix. The new version replaces /etc/searx/settings.yml and /etc/searx/uwsgi.ini
Build using "./manage.sh docker_build", add "push" as parameter also push the Docker image.
The script requires a git repository to work (it makes sure that the last git tag matches searx/version.py)
"git describe" is used to create a meaningful version.
Example : 0.15.0-90-49c5bcb4-dirty (dirty means that the docker image was made with uncommited changes).
Use "docker inspect -f {{.Config.Labels.version}} searx" to get the version of an existing image.
.dockerignore based on .gitignore
.travis.yml: include docker stage
