A deploy that exited before releasing its lock, such as after a failed build, left `.deploy.lock` behind, so `apps:locked` reported a lock nothing held and the build record stayed running. The lock is now released and the record finalized whenever the deploy exits. `apps:lock` never actually blocked deploys, as it only created the file deploys lock. It now creates its own lock that git pushes, `ps:rebuild`, `git:sync`, and the `git:from-*` commands refuse to deploy past, and `apps:locked` only reports a lock when the app was locked via `apps:lock` or a deploy is in progress. Locks created by `apps:lock` in older versions are migrated on install.
Plugin modules had drifted apart over time, leaving `github.com/onsi/gomega` pinned at `v1.43.0` in two plugins and `v1.43.1` in three others, with similar splits across a number of indirect dependencies. The `config` and `storage` plugins also carried manifests that were out of date enough that their Go tests refused to run, and the `builds` plugin was missing from the workspace entirely.
When the per-build log is missing, builds:output fell through to
journalctl without checking that the build record exists. A typo'd or
never-recorded id then printed nothing and exited 0.
Validate the build record first and return "no such build" when it is
absent. Existing records with pruned logs still use the journalctl
fallback.
Fixes#9031
cmd-git-sync calls dokku_setup_build_capture before the clone/fetch,
because it cannot know whether there is anything to deploy until it has
fetched. Both paths that then decline to build - a plain `git:sync` with
no build flag, and a `--build-if-changes` poll that finds the remote
unchanged - returned without ever finalizing that record.
The record stayed at status=running forever, which has three
consequences on an app polled on a timer:
- PruneAppBuilds only runs from builds-record-finalize, so nothing
prunes the ticks while the app is idle; they accumulate without bound,
invisibly, because builds:list caps output at the retention count.
- The next real deploy's ReapAbandonedBuilds finalizes each of them as
status=failed, exit_code=-1, with finished_at stamped at the moment of
reaping - indistinguishable on disk from a build that really failed,
and with a duration that grew to the polling interval.
- That same deploy's prune then keeps the newest records by started_at,
which after an idle stretch are the ticks, deleting the last real
deploy's record and log.
Add a builds-record-discard trigger that removes a non-terminal record
and its log, and call it from both no-op paths in cmd-git-sync. The
trigger leaves missing and already-terminal records alone, so it can
never delete real history.
Fixes#9030
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Every Go-implemented plugin's `:report --format json` now emits keys without the redundant `<plugin>-` head segment, matching the shape bash plugins have always emitted. The CLI flag names and `:set` semantics are unchanged. For backwards compatibility during the 0.38.x patch series, the legacy `<plugin>-<property>` keys are emitted side-by-side with the new keys, and a future major release will drop the legacy keys. `common.ReportSingleApp` is refactored to accept a `ReportSingleAppInput` struct with a `Validate()` method so the input is checked before any work runs, which also catches the latent `"docker options"` reportType bug at the API boundary.
Adds typed JSON build records under data/builds/<app>/<build-id>.{json,log} keyed on a stable base36 ULID-style DOKKU_BUILD_ID generated for every deploy. The new commands surface that history (builds:list, builds:info, builds:prune) and an operator-configurable retention via builds:set retention. The existing builds:cancel and builds:output now key on the build-id (with safe handling for already-finalized and abandoned records), and the per-build log file replaces journalctl as the durable source of truth for builds:output.
This still needs tightening to ensure we don't trigger this for normal git plugin commands - as well as anything that triggers a deploy - but works in a pinch.
The DOKKU_PID now never gets overwritten except in the case that DOKKU is executed by the sudo user. If the command ends up executing a deploy, then the pid of the `dokku` owned process - which may have been executed via sudo - will be written to the file lock, allowing future commands to interact with the original process.
Additionally, the new builds plugin can be used to handle killing a build.