-
Notifications
You must be signed in to change notification settings - Fork 61
[OPENJDK-4002] merge down jlink-dev branch (take 2) #596
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: ubi9
Are you sure you want to change the base?
Conversation
For push actions only (e.g., post-PR merge), after successfully completing the behave-test step, Log into GitHub container registry (ghcr.io) and push the image. Add the necessary permissions to write to the GHCR repository. Adjust the build and test steps to use a predictable, version-independent image name ("localimage") which we can reference in the new steps. Signed-off-by: Jonathan Dowland <[email protected]>
This allows GitHub to provide links back from the Container Images to the source repository that built them. See <https://github.com/opencontainers/image-spec/blob/main/annotations.md#pre-defined-annotation-keys> Signed-off-by: Jonathan Dowland <[email protected]>
…ontainer application
[OPENJDK-2063] Add dockerfile to do a multi-stage build for jlink JRE light weight container application
…nto run-java.sh - This tweak is to avoid defining JBOSS_CONTAINER_UTIL_LOGGING_MODULE as ENV in templates/jlink/Dokcerfile - Deleted run-env.sh and folded its content to run-java.sh as its not doing anything much other than setting value of JAVA_APP_DIR Signed-off-by: Jayashree Huttanagoudar <[email protected]>
Signed-off-by: Jayashree Huttanagoudar <[email protected]>
…GING_MODULE and JBOSS_CONTAINER_JAVA_RUN_MODULE Signed-off-by: Jayashree Huttanagoudar <[email protected]>
This is a hail-mary change: we are experiencing problems with GHA runs timing out (behave tests taking up to five hours to run). We were running on a not-current GHA runner version, and I think the old runners get fewer resources than the newer ones. Signed-off-by: Jonathan Dowland <[email protected]>
[OPENJDK-2063] Tweak run scripts to support docker multistage build for jlinked JRE container application
…nd also remove ENV setting as they are taken care in run-java.sh Signed-off-by: Jayashree Huttanagoudar <[email protected]>
[OPENJDK-2063] Modify Dockerfile to consolidate destination path for image scripts
Signed-off-by: Jayashree Huttanagoudar <[email protected]>
https://github.com/jboss-openshift/openshift-examples/ https://github.com/jboss-container-images/openjdk : tests/OPENJDK-1549 https://github.com/jerboaa/quarkus-quickstarts (branch quickstart-2.16-s2i-cds) https://github.com/jmtd/openshift-quickstarts (branch OPENJDK-631-quarkus-fast-jar) Signed-off-by: Jayashree Huttanagoudar <[email protected]>
Signed-off-by: Jayashree Huttanagoudar <[email protected]>
Signed-off-by: Jayashree Huttanagoudar <[email protected]>
Signed-off-by: Jayashree Huttanagoudar <[email protected]>
Signed-off-by: Jayashree Huttanagoudar <[email protected]>
[OPENJDK-2199] Update tests to refer new application source repository
Changes were made to run-java.sh in one branch of a merge that needed to be promoted up to the new pathfinder mdouel from the other branch of a merge. Signed-off-by: Jonathan Dowland <[email protected]>
Signed-off-by: Jonathan Dowland <[email protected]>
README for jlink template directory
Signed-off-by: Jonathan Dowland <[email protected]>
Signed-off-by: Jonathan Dowland <[email protected]>
Signed-off-by: Jonathan Dowland <[email protected]>
Signed-off-by: Jonathan Dowland <[email protected]>
Add JDK21 jmods RPM to builder image
For clarity. Signed-off-by: Jonathan Dowland <[email protected]>
[OPENJDK-3646] clean up jdeps *.txt files unless S2I_DELETE_SOURCE is set
This reverts commit 4967f2c. Should address OPENJDK-3680.
[OPENJDK-3680] drop gha stuff
[OPENJDK-3681] get rid of measure.sh
[OPENJDK-3682] Correct org.opencontainers.image.source
Align versions
Signed-off-by: Jonathan Dowland <[email protected]>
Signed-off-by: Jonathan Dowland <[email protected]>
Update the example APP_URI and CONTEXT_DIR
[OPENJDK-3686] update jlink README with RHEC link, customer focus
Signed-off-by: Jonathan Dowland <[email protected]>
Signed-off-by: Jonathan Dowland <[email protected]>
the s2i and jdk modules are used by the non-jlink images; remove the changes to those made in the jlink-dev work. Move those changes to the jlink module. Note: prior commit added an explicit dependency on the jlink module to the jlink-dev image descriptor. Signed-off-by: Jonathan Dowland <[email protected]>
Adjusting jlink-dev work to the expectations after OPENJDK-3655 Signed-off-by: Jonathan Dowland <[email protected]>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
While I understand the rationale, this seems a risky move. I worry that this will break the UBI 9 image scripts in ways we don't anticipate. By keeping it separate we only risk breakage on the UBI9 tech-preview image that uses the jlink work. That's a much smaller set of people than the full UBI9 set of images.
Should we instead merge this into the UBI10 branch?
A less-invasive change would be to introduce the pathfinder module, but not refactor the existing modules to use it. That would still leave changes (2) and (3) above. What's your assessment of the risk of those?
That's an interesting idea. At the moment, I've been updating and shipping the TP image alongside the UBI9 containers. Merging this to UBI10 would possibly prevent me doing that, but we would still get automated respins for CVEs based on the last shipped version. That's probably fine. |
https://issues.redhat.com/browse/OPENJDK-4002
This is a merge-down of the jlink-dev branch to ubi9. Rationale: to remove the need to continually sync the jlink-dev branch with ubi9 and ease rebuilding the TP image at the same time as the rest of them.
The jlink-dev branch originally modified some core modules and the JDK21 image descriptor. We now instead create a new image descriptor and all jlink changes are confined to that and the new jlink module, with the following exceptions:
modules/run/artifacts/run-java.sh
so it can be re-used elsewhere.modules/s2i/bash/artifacts/local/assemble
. These are entirely logic-gated byif [ "$S2I_ENABLE_JLINK" = "true" ]
, so none of it will be executed by the normal images.modules/s2i/core/artifacts/s2i-core
adds an export ofS2I_ENABLE_JLINK
(defaulting to false) which should be benign