16 Commits

Author SHA1 Message Date
Jose Diaz-Gonzalez
6836e56b24 refactor: add global docker build and run args
This reduces duplication and allows users to rely on global arguments to generate most labels needed.

The 'commit' command cannot have global arguments because the whitespace within the label change breaks string-based interpolation, and shell arrays cannot be exported.
2019-07-20 05:52:20 -04:00
Jose Diaz-Gonzalez
d1af10fc51 feat: add labels to all build, commit, create, and run docker calls
This will allow us to further filter containers and images by application when cleaning up containers.

Refs #3515
2019-07-19 15:45:43 -04:00
Jose Diaz-Gonzalez
785d9b4937 feat: allow setting DOCKER_BIN path for docker execution
This change allows operators to specify a DOCKER_BIN environment variable. This will specify a binary to run when executing docker, which is useful in cases where the 'docker' command being run must be modified in a way that would otherwise be invasive to Dokku, but minimalistic if done within a wrapper.
2019-05-29 00:46:19 -04:00
Jose Diaz-Gonzalez
243d5ecc43 feat: integrate resource plugin with scheduler-docker-local 2019-04-08 11:31:34 -04:00
Jose Diaz-Gonzalez
173b87928c feat: disable container restarts for stopped containers
This should avoid the case where the docker daemon starts the containers upon reboot.
2019-01-20 15:19:54 -05:00
Jose Diaz-Gonzalez
86795ddacc tests: run mvdan/shfmt on test runs
While I do not agree with _every_ style change, this will force Dokku to have consistent formatting across all shell scripts, which is arguably a Good Thing™.

The command used to reprocess everything is:

```shell
shfmt -l -bn -ci -i 2 -w .
```
2019-01-07 01:25:55 -05:00
Jose Diaz-Gonzalez
cf834f8ac8 fix: keep track of failed containers regardless of docker kill output
`docker kill` sometimes exits non-zero, meaning we don't keep track of logs for failed containers
2018-11-14 17:23:23 -05:00
Jose Diaz-Gonzalez
38ed9bff39 feat: add a way to retire old containers scheduled by the docker-local plugin
These will be killed on a cron schedule, allowing us to ensure old containers don't run once memory pressure clears up - even across restarts.

Closes #3225
2018-10-21 12:50:39 -04:00
Jose Diaz-Gonzalez
0d776ec2c8 fix: ensure we respect the herokuish start command 2018-10-11 20:19:56 -04:00
Jose Diaz-Gonzalez
0292c4ee3b refactor: move get_cmd_from_procfile to ps/functions
Procfile handling should only be performed in the ps plugin
2018-10-11 20:19:56 -04:00
Jose Diaz-Gonzalez
c924bdef04 feat: expand the PORT environment variable when provided
Refs #2793
Refs #3196
2018-10-11 20:19:56 -04:00
Jose Diaz-Gonzalez
78f467852e feat: add a subcommand for retrieving failed app deploy logs
This can be useful in cases where a developer wishes to see what error messages are available for a failed deploy, but does not wish to spelunk around `docker ps -a` output.

Note that the logs are not made available once a further deploy is triggered, or after the containers have been garbage collected.
2018-07-29 17:17:53 -04:00
Jose Diaz-Gonzalez
c9de100dd9 feat: add container id to retire-container-failed arguments (#3227)
Refs dokku/dokku#3225
2018-07-09 10:27:24 -05:00
Jose Diaz-Gonzalez
50a98db8e6 refactor: move container file cleanup/renames to scheduler-docker-local plugin 2018-04-06 22:53:02 -04:00
Jose Diaz-Gonzalez
331d58c62b fix: trigger must be executable 2018-04-06 22:53:02 -04:00
Jose Diaz-Gonzalez
5cada188d4 feat: move deploy logic into scheduler-docker-local plugin 2018-04-06 22:53:02 -04:00