Skip to content

CP-13653 linux-pkg: disable windows-connector's post-push auto-trigger, require manual trigger to publish - #416

Open
vimleshmishra wants to merge 1 commit into
developfrom
dlpx/pr/vimleshmishra/0ca1958f-2337-4dda-b9a7-86594d41bd00
Open

CP-13653 linux-pkg: disable windows-connector's post-push auto-trigger, require manual trigger to publish#416
vimleshmishra wants to merge 1 commit into
developfrom
dlpx/pr/vimleshmishra/0ca1958f-2337-4dda-b9a7-86594d41bd00

Conversation

@vimleshmishra

@vimleshmishra vimleshmishra commented Aug 27, 2026

Copy link
Copy Markdown
Contributor

Problem

windows-connector's post-push Jenkins job auto-triggers on every push to dlpx-app-gate, even though connector source changes are infrequent and dlpx-app-gate sees heavy, unrelated traffic. Most of these builds are pure rebuild churn with no actual connector source change, and any real build-health regression is indistinguishable from that noise without added machinery to tell the two apart.

Solution

Disable the auto-trigger and require a human to manually trigger a build/publish instead:

  • Set DEFAULT_PACKAGE_GIT_URL="none" in packages/windows-connector/config.sh. The shared Job DSL (devops-gate/jenkins/jobs/linux_pkg_build_package.groovy) already disables a job's push trigger when its git URL is falsy/"none"no devops-gate change is needed for this.
  • Since PACKAGE_GIT_URL would otherwise also resolve to "none" (breaking source fetch), override the fetch() hook in the same config.sh to hardcode the real dlpx-app-gate URL directly, bypassing that variable.
  • Add packages/windows-connector/README.md documenting the new process end-to-end: why the trigger was removed, how the mechanism works, how to bump connectorDebVersion, how to test via git-ab-pre-push -b windows-connector before merging, and how to publish a new latest after merging.

This supersedes the earlier automated skip-detection design (connectorDebVersion cache-check + NOT_BUILT signaling), tracked in #407 and devops-gate#4772 — both are being closed/discarded in favor of this simpler approach.

Testing Done

Verified end-to-end on a personal Jenkins developer controller (jenkins-vmishra.dcol2.delphix.com, via JenkinsManager configure -j ops -g linux-pkg pointed at this branch on my fork):

  • Trigger suppression: the regenerated windows-connector/post-push job has no SCM/push-trigger property, and its auto-generated description reads "Triggers are disabled for this job, so it must be run manually."
  • Source fetch still works: manually triggered post-push build Initial review #1 — it cloned the real dlpx-app-gate source via the fetch() override, BUILD SUCCESSFUL (Gradle), and the job finished SUCCESS end-to-end.
  • Publish path confirmed isolated: published to s3://snapshot-de-images/builds/jenkins-ops.vmishra/.../post-push/latest — the personal-controller-specific jenkins-ops.<user> S3 prefix, not real production latest.

Notes to Reviewers

Tracked under CP-13653, epic CP-13579.

A companion pointer comment was added in dlpx-app-gate's appliance/host/windows/build.gradle, next to connectorDebVersion, linking back to this package's README — since that's the actual entry point most people will hit when making a connector change.

@vimleshmishra
vimleshmishra force-pushed the dlpx/pr/vimleshmishra/0ca1958f-2337-4dda-b9a7-86594d41bd00 branch from 7ff43fd to 158cda0 Compare August 27, 2026 10:50
@vimleshmishra
vimleshmishra force-pushed the dlpx/pr/vimleshmishra/0ca1958f-2337-4dda-b9a7-86594d41bd00 branch from 158cda0 to 761aa96 Compare August 27, 2026 10:53
@vimleshmishra
vimleshmishra requested review from prakashsurya and sebroy and removed request for sebroy August 27, 2026 10:55
@vimleshmishra
vimleshmishra marked this pull request as ready for review August 27, 2026 10:57
Comment on lines +65 to +71
## Known caveat

`GIT_HASH`/`BUILD_INFO` published alongside the artifact reflect the commit at
the time of the last *real* build, not necessarily current `develop` HEAD —
this is true regardless of whether a build was triggered automatically or
manually, and is a pre-existing limitation, not something this change
introduces.

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm not sure I agree with the framing that this is a "caveat".. I view this as a feature, and working as intended..

I'm OK leaving it, just makes it seem like this is a defect, when it's actually the whole point of capturing that information at build time.. so we can map the package contents, to the source code that was used to generate it..

which, by design, will lag the app-gate develop branch for this package..

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Development

Successfully merging this pull request may close these issues.

2 participants