Closed Bug 1131133 Opened 10 years ago Closed 10 years ago

The UI sometimes creates a duplicate instead of updating a job

Categories

(Tree Management :: Treeherder, defect, P1)

defect

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: mdoglio, Assigned: mdoglio)

References

Details

Attachments

(2 files)

Recent changes introduced a regression in the job updates: sometimes a new job button is created instead of updating an existing one
Summary: The UI sometimes creates a duplicate instead of updating a jobs → The UI sometimes creates a duplicate instead of updating a job
Blocks: 1097090
Priority: -- → P1
Attachment #8564232 - Flags: review?(cdawson)
Attachment #8564230 - Flags: review?(cdawson)
Commit pushed to master at https://github.com/mozilla/treeherder-service https://github.com/mozilla/treeherder-service/commit/dbbf579b9bb9a0def04cdb849b287fabcf2e2356 Bug 1131133 - group jobs by id This is a quick&dirty solution to circumvent the problem of the duplicate signatures. The buildername is selected randomly by the group by, but that's more or less what we have done until now so we can probably keep living with that.
Commit pushed to master at https://github.com/mozilla/treeherder-ui https://github.com/mozilla/treeherder-ui/commit/758880230ac92c2ec11fa834bdf42b09c596cb4d Bug 1131133 - fix regression on job updates The job poller was using a timestamp generated on the client to retrieve the updates. Under some circumstances (eg a significant db replication lag) the client was missing some updates. Now it's using the latest last_modified fetched. If for some reason we miss an update we re-run the same request on the next iteration
Attachment #8564230 - Flags: review?(cdawson) → review+
Attachment #8564232 - Flags: review?(cdawson) → review+
Assignee: nobody → mdoglio
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
Commit pushed to master at https://github.com/mozilla/treeherder https://github.com/mozilla/treeherder/commit/7c680de2b99b96a60f227f075ea76ead442f5b13 Bug 1131133 - fix regression on job updates The job poller was using a timestamp generated on the client to retrieve the updates. Under some circumstances (eg a significant db replication lag) the client was missing some updates. Now it's using the latest last_modified fetched. If for some reason we miss an update we re-run the same request on the next iteration
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: