Investigate publish recovery after an interrupted index sync #102

New Issue

2026-10-05T12:50:13Z

eleith commented

2026-10-05 12:50:13 +00:00

Goal

Investigate how an interrupted publish recovers before choosing any implementation change.

Potential failure window: core index/cache updates succeed inside luzzle.sync, but execution stops before the durable step result is recorded. A retry may then see an empty diff despite unfinished downstream publication.

Investigation

Reproduce the suspected interruption with isolated fixtures and controlled failure injection.
Distinguish replay after a completed checkpoint from retry before that checkpoint, and from a fresh publish after failure.
Observe which pieces still reach web sync and asset generation, including relevant pruning behavior.
Report what existing OpenWorkflow recovery already handles and any demonstrated gap.
Review findings and options with the owner before changing behavior.

No work manifest, new persistent state, retry policy, or automatic full rebuild is preselected. Individual transform failures remain best-effort publication with a warning summary, not a reason to abort the publish.

## Goal Investigate how an interrupted publish recovers before choosing any implementation change. Potential failure window: core index/cache updates succeed inside `luzzle.sync`, but execution stops before the durable step result is recorded. A retry may then see an empty diff despite unfinished downstream publication. ## Investigation - Reproduce the suspected interruption with isolated fixtures and controlled failure injection. - Distinguish replay after a completed checkpoint from retry before that checkpoint, and from a fresh publish after failure. - Observe which pieces still reach web sync and asset generation, including relevant pruning behavior. - Report what existing OpenWorkflow recovery already handles and any demonstrated gap. - Review findings and options with the owner before changing behavior. No work manifest, new persistent state, retry policy, or automatic full rebuild is preselected. Individual transform failures remain best-effort publication with a warning summary, not a reason to abort the publish.

eleith added the bug label 2026-10-05 12:50:13 +00:00

eleith changed title from ~~Retain publish work across an interrupted index-sync checkpoint~~ to Investigate publish recovery after an interrupted index sync

2026-10-05 23:14:51 +00:00

eleith referenced this issue

2026-10-06 14:55:04 +00:00

Report publish transform failures honestly and keep healthy pieces processing #100

eleith referenced this issue

2026-10-06 19:33:37 +00:00

Report publish transform failures and continue healthy pieces #109

eleith referenced a pull request that will close this issue

2026-10-06 23:13:09 +00:00

Recover unfinished publishes using content hashes #110

eleith closed this issue

2026-10-06 23:23:31 +00:00