since the introduction of the php base image version information in the webapp container are doubled. by removing the build args from the php image at least there is no "newest" version any longer for webapp, but it still shows the version twice.
this introduces a hotfix for this behaviour.
Signed-off-by: Felix Bartels <felix@host-consultants.de>
* adapt webapp container for new php base image
* clean out webapp dockerfile
* complete switch of webapp to php-fpm
* update config in web container
* update readme
* remove logging config from docker-compose.yml
* add php container to tagging and publishing
* add docker pull to warm up build cache on travis
* explicitly build multi stage build stages
* make create repo script dynamic in the choice of the distribution
* change current config changes to use php_cfg_gen
* add remaining (for kopano relevant config files)
* add autodiscover to z-push image
* add more autodiscover aliases to kweb
fixes#114#39
* add scheduler container for gabsync
Adds a general scheduler container to trigger tasks within containers.
Also adds gabsync to zpush image
* make scheduler dynamic
execute each cron job once at startup to see if they would succeed
* remove services scripts from core container (now is part of the scheduler)
* add "CRONDELAY" for tasks that should not be executed at startup
* add documentation
* clear out crontab at startup
* move .env creation into install
rerun check-scripts after creation as well
* update docker-compose
* add simple test if containers start up
* redirect output of compose stop
* Add dynamic dependency-wait for server and spooler service.
Instead of using hard-coded dependency-waits on startup for database and MTA for the server and spooler, respectively, the actually configured values are now used.
For the servers, this also takes the use of a unix file socket over a network socket into account.
* add example docker override (one which would disable the mail container)
* move some config options from docker-compose into the default conf
* add missing depends_on
* Prefix container names with COMPOSE_PROJECT_NAME.
This way it is easily possible to use several instances of this service stack on one host via different .env-files (e.g. for staging/production).
Volumes and networks are separated in this manner by docker-compose anyway, now the containers are too.
* Use underscore in container-names for consistency with networks and volumes.
* Remove "kopano" from container names.
This avoids repetition in the names when using the default project name which also "kopano".
* Make Caddy default port configurable.
* get settings for meet from env
* prepare ssl container for device registration for konnect/kwmserver
* move device registry modification to konnect container
* enable WebApp in the app switcher
* upstream docker container has been updated to alpine 3.9
* update konnect
* add further config for guest mode
* replace the check for the file with a check for konnect startup
* fix kwmserver wrapper
* add possibility to change logging in kwmserver
* add hadolint for dockerfile linting
* add hadofile config
* add checks for dockerfiles and shellcheck into makefile
* shellcheck fixes
* add workaround so that .env can be sourced again from version.sh
* hadolint fixes
* print progress of build/run.sh
* fix check for jq in setup.sh
relates to #41 and #26
I try to integrate the backup mechanisms of https://github.com/blacklabelops/volumerize, handling named volumes and automatic service stop and restart for consistent backups. Nevertheless this needs a clean stop of a service, which is only able to be achieved, if the docker restarts the containers only if they are not explicitly stopped.
