Chart 2026.7.1: keep an execution-capable member in the default queue - #4
Merged
Conversation
register_queue only assigns instances when it creates the group. On an upgrade the group already exists, so it prints "Instance Group already registered" and leaves it empty -- and an empty regular instance group accepts job launches and never runs them. Assert membership here rather than trusting that call, and fold the separate node_type fix-up into the same block: the backend now honours FORAIL_NODE_TYPE when the task pod registers itself, but a newer chart can still meet an older image, and this Job is what makes the two converge.
Closes the changelog section that still said [Unreleased] while 2026.7.0 was already out.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The init Job now asserts membership of the
defaultinstance group instead of trustingregister_queue, which assigns instances only when it creates the group — so on an upgrade it printedInstance Group already registered defaultand left the group empty. An empty regular instance group accepts job launches and never runs them.Pins
images.backend.tagto2026.07.1, which carries the matching backend fix (the task pod used to re-register itself ascontroland flipdefaultback to a ContainerGroup on every start). The frontend stays at2026.07.0— unchanged.Verified on a fresh 7-node cluster: install 2026.06.0 → upgrade to this chart → job successful with no manual intervention, and the state survived two task-pod restarts. Full Cypress regression against that upgraded install: 102 tests, 101 passing, 0 failing.