fix(smoke): treat UPDATE_ROLLBACK_COMPLETE as needs-recreate#329
Merged
Conversation
…loyable Post-merge smoke on main found all-demo in UPDATE_ROLLBACK_COMPLETE, called use_canonical, and the subsequent CFN update failed because the internal rollback (triggered by a leaf failure) tried to delete the StorageFileSystem61EA7B3D — which has pending S3 export data and needs forceDelete=true. The AWS::S3Files::FileSystem CFN handler does not pass forceDelete, so the rollback fails and the deploy reports "Failed to create/update the stack". CFN technically accepts updates from UPDATE_ROLLBACK_COMPLETE, but for the all-demo umbrella that state always hides this kind of half-cleaned S3Files / nested-stack debris. Safer to mirror the ROLLBACK_COMPLETE branch: delete-stack, wait, retain-on-DELETE_FAILED, then `continue` the loop so the next iteration hits DOES_NOT_EXIST → use_canonical → full resource sweep (which DOES force-delete file systems). Cost: one extra ~60m recreate cycle when CFN rolled back. Benefit: the umbrella self-recovers from S3Files-stuck rollbacks instead of needing human cleanup.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Post-merge smoke on main found `all-demo` in `UPDATE_ROLLBACK_COMPLETE` and called `use_canonical` (because CFN technically accepts updates from that state). The subsequent CFN deploy then failed:
```
StorageFileSystem61EA7B3D DELETE_FAILED:
File system 'fs-...' has data pending export to S3.
Use forceDelete=true to force delete without exporting pending data.
```
The leaf update failed, CFN's internal rollback then tried to delete the `AWS::S3Files::FileSystem` — whose CFN handler does not pass `forceDelete=true`. Update fails. The umbrella ends up wedged again.
Fix
Mirror the `ROLLBACK_COMPLETE` branch: nuke from `UPDATE_ROLLBACK_COMPLETE` and let the loop re-evaluate. Next iteration hits `DOES_NOT_EXIST` → `use_canonical` → full resource sweep (which does call `s3files delete-file-system --force-delete`), then fresh deploy.
Costs an extra ~60 min recreate cycle each time CFN rolled back; benefit is the umbrella self-recovers from S3Files-stuck rollbacks rather than needing human intervention.
Test plan