The Redis-based JobQueue had a feature to abandon jobs that were sitting in the queue for longer the claimTTL time period, we need to have a similar feature ChangeProp. Also, if some enormous time period has passed from the root job being submitted, like a week, I assume it's safe to reject the jobs generated by that root jobs. This will probably be a no-op in the new queue cause the wait periods should not be that high, but it would help clear the backlog carried over from the old queue when switching jobs.
Description
Description
Details
Details
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | • Pchelolo | T157088 [EPIC] Develop a JobQueue backend based on EventBus | |||
| Resolved | • Pchelolo | T183744 FY17/18 Q3 Program 8 Services Goal: Migrate two high-traffic jobs over to EventBus | |||
| Resolved | • Pchelolo | T189303 Support claimTTL and rootClaimTTL in change-prop |
Event Timeline
Comment Actions
Change 417946 had a related patch set uploaded (by Ppchelko; owner: Ppchelko):
[mediawiki/services/change-propagation/jobqueue-deploy@master] Abandon refreshLinks if the root job is more then 1 week long.
Comment Actions
Change 417946 merged by Mobrovac:
[mediawiki/services/change-propagation/jobqueue-deploy@master] Abandon refreshLinks if the root job is more then 1 week long.
Comment Actions
Mentioned in SAL (#wikimedia-operations) [2018-03-09T15:59:07Z] <ppchelko@tin> Started deploy [cpjobqueue/deploy@5795526]: Enable root_claim_ttl for refreshLinks T189303
Comment Actions
Mentioned in SAL (#wikimedia-operations) [2018-03-09T15:59:46Z] <ppchelko@tin> Finished deploy [cpjobqueue/deploy@5795526]: Enable root_claim_ttl for refreshLinks T189303 (duration: 00m 38s)