We've seen a few dozen lock failures since the upgrade. This might be caused by dedupe queries, or by something more mysterious.
Description
Description
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | Eileenmcnaughton | T139973 CiviCRM db lock damaging queue consumption | |||
| Resolved | Eileenmcnaughton | T139964 CiviCRM dedupe causes runaway query | |||
| Declined | None | T140694 Don't reject messages on unknown database error |
Event Timeline
Comment Actions
Thanks. We still have to make the consumer more robust to timeouts, T116493 and the consumer should retry a message rather than rejecting it, but that's followup work...