[DOP-25451] Fix Hive and HDFS connector ignoring user & password

Author: dolfinus

docker/Dockerfile.worker

@@ -66,5 +66,13 @@ RUN --mount=type=cache,target=/root/.cache/pypoetry \
         --without docs,dev \
     && python -m compileall -j 4 .venv
 
-RUN sed -i 's/python -m/coverage run -m/g' /app/entrypoint.sh
 ENV SYNCMASTER__WORKER__CREATE_SPARK_SESSION_FUNCTION=tests.spark.get_worker_spark_session
+
+# Collect coverage from worker
+RUN sed -i 's/python -m/coverage run -m/g' /app/entrypoint.sh
+
+# Replace kinit binary with dummy, to skip Kerberos interaction in tests
+RUN mkdir -p /app/.local/bin && \
+    echo "#!/bin/bash" > /app/.local/bin/kinit \
+    && chmod +x /app/.local/bin/kinit
+ENV PATH="/app/.local/bin:$PATH"

docs/changelog/0.2.2.rst

@@ -0,0 +1,8 @@
+0.2.2 (2025-04-11)
+==================
+
+Bug fixes
+---------
+
+- Call ``kinit`` before starting Spark session conecting to ``Hive`` cluster.
+- Fix ``HDFS`` connection was trying to use anonymous auth instead of user/password.

docs/changelog/index.rst

@@ -3,6 +3,7 @@
     :caption: Changelog
 
     DRAFT
+    0.2.2
     0.2.1
     0.2.0
     0.1.5

syncmaster/worker/handlers/file/hdfs.py

@@ -25,4 +25,6 @@ def connect(self, spark: SparkSession):
 
         self.file_connection = HDFS(
             cluster=self.connection_dto.cluster,
+            user=self.connection_dto.user,
+            password=self.connection_dto.password,
         ).check()

syncmaster/worker/spark.py

@@ -6,8 +6,14 @@
 import math
 from typing import TYPE_CHECKING
 
+from onetl.connection.kerberos_helpers import kinit_password
+
 from syncmaster.db.models import Run
-from syncmaster.dto.connections import ConnectionDTO
+from syncmaster.dto.connections import (
+    ConnectionDTO,
+    HDFSConnectionDTO,
+    HiveConnectionDTO,
+)
 
 if TYPE_CHECKING:
     from pyspark.sql import SparkSession
@@ -33,6 +39,10 @@ def get_worker_spark_session(
         log.debug("Enabling Hive support")
         spark_builder = spark_builder.enableHiveSupport()
 
+    for entity in source, target:  # type: ignore
+        if isinstance(entity, (HiveConnectionDTO, HDFSConnectionDTO)) and entity.user and entity.password:  # type: ignore
+            kinit_password(entity.user, entity.password)
+
     return spark_builder.getOrCreate()