Keep a crashed browser from eating the CI job - #668
Merged
Conversation
A job intermittently produces no output after a build until the runtime cap cancels it, leaving chromedriver, chrome, and two crashpad handlers as orphans: the renderer crashes and the driver waits on it forever. Run Chrome with `--disable-dev-shm-usage`, keeping it off the small `/dev/shm` of CI runners and containers, where exhausting it crashes the renderer. As a second line of defense, cap `:js` examples at five minutes, so a hung browser fails one example with a backtrace and lets the rest of the suite run, instead of silently consuming the job's runtime cap. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
CI jobs intermittently go silent right after an
ember buildfinishes, untiltimeout-minutes: 15cancels them (e.g. this job; the same signature cancelled a job this morning). The orphan-process list at cleanup showschromedriver,chrome, and twochrome_crashpad_handlerprocesses: the renderer crashes mid-example and the driver waits on it forever, so the job produces no output and no failure.--disable-dev-shm-usage. Chrome's shared memory lives in/dev/shm, which is small on CI runners and containers; exhausting it crashes the renderer — the crashpad handlers are its trace. This is the standard mitigation for headless Chrome in CI.:jsexamples at five minutes (theTimeout.timeout(300)arrangement theApp#testexample already uses). A hung browser then fails one example with a backtrace — and the existing failure hook dumps the page and browser console — and the rest of the suite still runs, instead of the job being cancelled with no diagnostics.Verification
js: falsefeature scenarios pass locally with the reconfigured support file loaded (the:jsexamples need Chrome, which CI provides)/dev/shmmechanism match, and the timeout layer bounds any residual hang