I believe the troublesome tasks are these three: https://github.com/apache/beam/blob/2592c8396ae1a68dadc351135c2a79397fb70942/website/build.gradle#L95-L108 (to avoid confusion, note that these are being added as part of the PR, although they appear in a commit on apache/beam in the link).
Nam, can you change the gradle tasks to avoid using -u root in docker? On Wed, May 6, 2020 at 1:24 PM Brian Hulette <bhule...@google.com> wrote: > I added some detail about the failures in the PR [1]. It seems for some > reason this PR is causing those directories (node_modules and themes) to be > created with root ownership. All subsequent runs of those jobs on the same > worker (for other PRs as well) try to clean up the old workspace, and they > fail with permissions errors. It looks like maybe +Udi Meiri > <eh...@google.com> debugged something similar recently in [2]? > > [1] https://github.com/apache/beam/pull/11554#issuecomment-624346035 > [2] https://issues.apache.org/jira/browse/BEAM-9737 > > On Wed, May 6, 2020 at 1:01 PM Aizhamal Nurmamat kyzy <aizha...@apache.org> > wrote: > >> Hi all, >> >> Nam and Brian have been working together on blog post names and >> the decision was to keep them as they are for now, because Hugo doesn’t >> fully support the feature that was mentioned earlier [1]. Also I believe >> this can be done after merging the PR. >> >> 1) Currently, the main blocker for merging is Staging Test Failures. >> Michal showed Nam how to handle the 1st test which was about Apache License >> missing. >> >> However, the 2nd and 3rd tests looked like some kind of permissions error >> on the Jenkins worker, not to be configured by code. For more details based >> on Jenkin logs, the 2nd test failed because of website/www/site/themes and >> the 3rd test failed because of website/www/node_modules, they are both >> auto-generated files on build. Can someone help Nam to look into this? >> >> RAT ("Run RAT PreCommit") — FAILURE >> Website_Stage_GCS ("Run Website_Stage_GCS PreCommit") — FAILURE >> Website_Stage_GCS ("Run Website_Stage_GCS PreCommit") — FAILURE >> >> 2) Are there any other blockers for merging? @Ahmet/Robert/others please >> share if there are any other blockers. >> >> >> [1] https://github.com/gohugoio/hugo/pull/4494 >> >> >> On Wed, May 6, 2020 at 10:19 AM Robert Bradshaw <rober...@google.com> >> wrote: >> >>> On Mon, May 4, 2020 at 7:07 PM Ahmet Altay <al...@google.com> wrote: >>> > >>> >> On Mon, May 4, 2020 at 6:30 PM Robert Bradshaw <rober...@google.com> >>> wrote: >>> >>> >>> >>> I took the massive commit and split it up into: >>> >>> >>> >>> (1) Infrastructure changes (basically everything outside of >>> >>> (website/www/site/content) >>> >>> (2) Sed script changes, and >>> >>> (3) Manual changes (everything not in (1) and (2)). >>> > >>> > >>> > Thank you Robert. This makes it much easier. What is the source of the >>> sed script? I am not sure why some of those lines are there. It would be >>> much easier for us to comment on the script source if it is reviewable >>> somewhere. >>> >>> I just gathered up common patterns as I was trying to go through and >>> review the files... Mostly it was an exercise in finding a compact >>> representation for the delta, not trying to be a perfect conversion. >>> (I do think in retrospect, if we do something like this again, it >>> would be preferable to commit a script that does the auto-conversion >>> (maybe even with some patch files for manual changes) both for ease of >>> reviewing and to avoid the stop-the-world situation we're in now. (I'm >>> still worried that some changes will get lost in the shuffle.) >>> >>