Closed Bug 1406972 Opened 8 years ago Closed 8 years ago

Use EMR 5.8.0/Spark 2.2 as the default for all airflow tasks

Categories

(Data Platform and Tools :: General, enhancement, P2)

enhancement
Points:
2

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: bugzilla, Unassigned)

Details

Attachments

(1 file)

Pinning to EMR 5.8.0 reduced the duration of the Experiment Aggregates job by nearly 1/3, and longitudinal runtimes also improved, so it seems it's stable and we might see significant improvement in the rest of our jobs. We should run trials for our p1 datasets (main_summary, crash_summary, churn, telemetry_aggregates are probably a good place to start?) and then change the default.
A little bit late to the party, but this PR bumped the version to 5.9.0.
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
Component: Scheduling → General
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: