Fix #126: use real user id in batch jobs instead of proxy user id

openEO processing in batch jobs needs real user id
for `load_result` and UDPs to work properly.
Proxy user id is just necessary when submitting the job.

openeogeotrellis/deploy/batch_job.py

@@ -239,7 +239,9 @@ def main(argv: List[str]) -> None:
     metadata_file = job_dir / argv[5]
     api_version = argv[6] if len(argv) >= 7 else None
     dependencies = _deserialize_dependencies(argv[7]) if len(argv) >= 8 else {}
-    user_id = argv[8] if len(argv) >= 9 else None
+    # TODO: do we still need proxy user id inside the batch job itself?
+    proxy_user_id = argv[8] if len(argv) >= 9 else None
+    user_id = argv[9] if len(argv) >= 10 else None
 
     _create_job_dir(job_dir)
 

openeogeotrellis/deploy/submit_batch_job.sh

@@ -134,4 +134,4 @@ spark-submit \
  --conf spark.yarn.tags=openeo \
  --jars "${extensions}","${backend_assembly}" \
  --name "${jobName}" \
- "${main_py_file}" "$(basename "${processGraphFile}")" "${outputDir}" "${outputFileName}" "${userLogFileName}" "${metadataFileName}" "${apiVersion}" "${dependencies}" "${proxyUser}"
+ "${main_py_file}" "$(basename "${processGraphFile}")" "${outputDir}" "${outputFileName}" "${userLogFileName}" "${metadataFileName}" "${apiVersion}" "${dependencies}" "${proxyUser}" "${userId}"

openeogeotrellis/deploy/submit_batch_job_spark24.sh

@@ -144,4 +144,4 @@ spark-submit \
  --conf spark.yarn.tags=openeo \
  --jars "${extensions}","${backend_assembly}" \
  --name "${jobName}" \
- "${main_py_file}" "$(basename "${processGraphFile}")" "${outputDir}" "${outputFileName}" "${userLogFileName}" "${metadataFileName}" "${apiVersion}" "${dependencies}" "${proxyUser}"
+ "${main_py_file}" "$(basename "${processGraphFile}")" "${outputDir}" "${outputFileName}" "${userLogFileName}" "${metadataFileName}" "${apiVersion}" "${dependencies}" "${proxyUser}" "${userId}"

openeogeotrellis/deploy/submit_batch_job_spark3.sh

@@ -165,4 +165,4 @@ spark-submit \
  --conf spark.yarn.tags=openeo \
  --jars "${extensions}","${backend_assembly}" \
  --name "${jobName}" \
- "${main_py_file}" "$(basename "${processGraphFile}")" "${outputDir}" "${outputFileName}" "${userLogFileName}" "${metadataFileName}" "${apiVersion}" "${dependencies}" "${proxyUser}"
+ "${main_py_file}" "$(basename "${processGraphFile}")" "${outputDir}" "${outputFileName}" "${userLogFileName}" "${metadataFileName}" "${apiVersion}" "${dependencies}" "${proxyUser}" "${userId}"