The periodic `dokku-retire` timer could run `ps:retire` between the test's final deploy and its own `ps:retire` call, processing the retire list first and leaving the test with no output to assert on.
A failed deploy could leave a non-empty extraction directory behind, such as the scheduler-k3s `kustomization` directory, which caused every following deploy to fail. Leftovers from previous deploys are now fully removed, removing a directory such as `config/kustomize` from the repository now removes the stored copy, and openresty includes now persist across restarts.
deploy_app cannot push onto an app that already has git history from git:from-image. git:from-archive deploys and clears source-image the same way git_4 already does.
Image retirement removes images by id, so a queued image could be deleted even though another app had since started using it as its git:from-image source image. Skip such images at retire time and drop them from the retire list.
The git_4 test images are also given a per-app label so they no longer share a cached image id, which let a retire entry left by one test delete the next test's source image.
Procfile commands for the `docker-local` scheduler were split differently on deploy and on `dokku run`. Deploys failed on escaped quotes, and `dokku run` ignored quotes entirely. Both now split commands using shell quoting rules, and commands containing bare shell operators fail with a clear error, matching the `k3s` scheduler.
The `<plugin>-global-<property>` keys in `:report` output returned the resolved value with the built-in default substituted in, so external tooling could not distinguish a property that had been set globally to the default from one that had never been set. The bare keys for caddy, haproxy, and traefik global-only properties had the same shape. The `global-<property>` keys now hold the raw stored value and are empty when nothing has been set, and a new `computed-<property>` key holds the effective value used at runtime, falling back through the per-app value (where one exists), the global value, and the built-in default. The bare global-only keys for the three proxy plugins are removed in favor of the raw/computed pair. Closes#8631.
The bare `init-process` and `parallel-schedule-count` keys previously returned the computed value, so external tooling could not tell whether a property had been set on the app or was merely defaulting. Both properties are now also configurable with `--global`, the report exposes `computed-*` and `global-*` keys alongside the bare raw keys, and the deploy path honors the global value before falling back to the linuxserver.io vendor heuristic.
Several plugin `:report` subcommands erroneously failed with `not deployed` when an info-flag matched a property that was empty. Empty values are legitimate for configuration properties pre-deploy and the `--format json` path already returns them without error. Remove the `value_exists` check across nginx, checks, git, certs, scheduler-docker-local, and the builder-* plugins so the info-flag form behaves consistently with the JSON form.
The argument-handling block in each plugin's report command overwrote APP with --global before checking whether $1 actually held an info flag, so `dokku <plugin>:report --global --<plugin>-foo` always emitted the full report instead of the requested value. Reorder the conditionals across the sixteen plugins that share the pattern so the info-flag check runs first, and add a regression bats case for each one.
The previous assertion piped grep -c through `|| echo 0`, which printed an extra `0` when the file existed without matches because grep -c emits the count of `0` and exits 1, triggering the fallback.
When ps:rebuild runs against an image-based deploy via git:from-image, the resulting image often shares the same SHA as the previous deployment, so retiring the old container's image would target the live image of the new container. The retirement is now skipped when another running container of the same app still references the image, and the cron retire loop self-heals previously stuck entries the next time it encounters them.
The docker-local scheduler now sends `SIGTERM` to old containers immediately after a successful deploy via `docker container kill --signal=SIGTERM`, rather than waiting `wait-to-retire` seconds before signaling. This matches Heroku's graceful-shutdown contract and lets applications begin draining in-flight work as soon as proxy traffic switches. The existing `wait-to-retire` grace period and `stop-timeout-seconds` hard-stop continue to apply unchanged as the authoritative cleanup path.
The intermediate container count check raced with the existing
zero-downtime retire of the previous container, so docker ps still
saw the renamed-but-pending container until wait-to-retire elapsed.
When scaling a process type to a smaller count, the indices above the
new count had their `CONTAINER.<proctype>.<idx>` state files deleted but
the underlying Docker containers were never registered for retirement,
so they remained running indefinitely. The cleanup block now reads each
orphaned state file and registers its container id with the existing
`scheduler-register-retired` trigger before the file is removed.
When developing Dokku, it is useful to use a real domain to test actual functionality. This is impossible if the domain is hardcoded to dokku.me in testing.
This allows folks to deploy apps that don't have a web process without needing to scale that process down before/after the first deploy. Note that the formations key in the app.json or a manual scale of other processes will be necessary to start anything non-web.
Closes#5700
In cases where a docker arg uses backticks - which may happen with docker labels - scheduled containers would fail to support these due to performing a double-eval. This is particularly noticeable for traefik labels, where backticks and parenthesis are common. This fix properly handles docker args as an array instead of performing an extra eval when scheduling containers.