Known Issues in Spark

Learn about the known issues in Spark, the impact or changes to the functionality, and the workaround.

CDPD-60862: Rolling restart fails during ZDU when DDL operations are in progress

During a Zero Downtime Upgrade (ZDU), the rolling restart of services that support Data Definition Language (DDL) statements might fail if DDL operations are in progress during the upgrade. As a result, ensure that you do not run DDL statements during ZDU.

The following services support DDL statements:
  • Impala
  • Hive – using HiveQL
  • Spark – using SparkSQL
  • HBase
  • Phoenix
  • Kafka

Data Manipulation Lanaguage (DML) statements are not impacted and can be used during ZDU. Following the successful upgrade, you can resume running DDL statements.

None. Cloudera recommends modifying applications to not use DDL statements for the duration of the upgrade. If the upgrade is already in progress, and you have experienced a service failure, you can remove the DDLs in-flight and resume the upgrade from the point of failure.
CDPD-67517: Spark3 tests are failing if /tmp is mounted as noexec.
Map the tmpdir to a writable path in spark3-conf/spark-defaults.conf using the following steps:
  1. Go to the Cloudera Manager.
  2. Go to the Spark service.
  3. Click the Configuration tab.
  4. Select Scope > Gateway.
  5. Select Category > Advanced.
  6. Locate the Spark Client Advanced Configuration Snippet (Safety Valve) for spark3-conf/spark-defaults.conf_client_config_safety_valve property.
  7. Map the tmpdir to a writable path:
    spark.driver.extraJavaOptions=-Djava.io.tmpdir=/var/tmp
                    spark.executor.extraJavaOptions=-Djava.io.tmpdir=/var/tmp
  8. Enter a Reason for change, and then click Save Changes to commit the changes.
  9. Deploy the client configuration.
CDPD-23817: In the upgraded Cluster, the permission of /tmp/spark is restricted due to the HDP configuration hive.exec.scratchdir=/tmp/spark.
If you are using the /tmp/spark directory in the CDP cluster, you must provide the required additional Policies/ACL permissions.
CDPD-22670 and CDPD-23103: There are two configurations in Spark, "Atlas dependency" and "spark_lineage_enabled", which are conflicted. The issue is when Atlas dependency is turned off but spark_lineage_enabled is turned on.
Run Spark application, Spark will log some error message and cannot continue. That can be restored by correcting the configurations and restarting Spark component with distributing client configurations.
CDPD-23007: Mismatch in the Spark Default DB Location. In HDP 3.1.5, hive_db entities have one attribute - 'location' which is configured to the '/managed' path. In fresh install of CDP 7.1.7, hive_db entities now have 2 attributes 'location' configured to '/external' path and 'managedLocation' configured to '/managed' path. In. the AM2CM migration (HDP 3.1.5 -> CDP 7.1.7), the 'location' attribute from hive_db entities in HDP 3.1.5 comes unaltered to CDP 7.1.7 and hence maps to '/managed' path.
This issue arises only if you are upgrading from HDP 3.1.5 to CDP 7.1.7. If you are performing a fresh install of CDP 7.1.7, you can ignore this issue.
None
CDPD-217: The Apache Spark connector is not supported
The old Apache Spark - Apache HBase Connector (shc) is not supported in CDP releases.
Use the new HBase-Spark connector shipped in CDP release.
CDPD-3038: Launching pyspark displays several HiveConf warning messages
When pyspark starts, several Hive configuration warning messages are displayed, similar to the following:
23/08/02 08:37:26 WARN conf.HiveConf: HiveConf of name hive.metastore.runworker.in does not exist
            23/08/02 08:37:26 WARN conf.HiveConf: HiveConf of name hive.masking.algo does not exist
            23/08/02 08:37:34 WARN conf.HiveConf: HiveConf of name hive.metastore.runworker.in does not exist
            23/08/02 08:37:34 WARN conf.HiveConf: HiveConf of name hive.masking.algo does not exist
These errors can be safely ignored.