search

Open-EO / openeo-geopyspark-driver

OpenEO driver for GeoPySpark (Geotrellis)
Apache License 2.0
26 stars 4 forks source link

batch jobs: avoid storing process graph twice #658

Open bossie opened 9 months ago

bossie commented 9 months ago

Less relevant in the context of dropping ZK in favor of ES but a batch job's process graph is being stored twice: once in the job, and a second time in the job result metadata.

soxofaan commented 9 months ago

sidenote: technically speaking, the openEO API allows updating the process graph (PATCH /jobs/{job_id}), so there could be a technical reason to keep both.

Personally I'm not a fan of this kind of mutability of batch job metadata, and tried to push against it, but it's still in the API spec.

bossie commented 9 months ago

The OpenEO API allows it but until we actually start supporting mutation, we're just wasting space.

Minor issue nonetheless.