Hive 2.3.3-1808 (EEP 6.0.0) Release Notes
This section provides reference information, including new features, fixes, known issues, and limitations for Hive 2.3.3-1808 EEP 6.0.0.
These release notes contain only MapR-specific information and are not necessarily cumulative in nature. For information about how to use the release notes, see Ecosystem Component Release Notes.
Hive Version | 2.3.3 |
Release Date | September 2018 |
MapR Version Interoperability | See Hive and HCatalog Support Matrix and Ecosystem Support Matrix (Pre-5.2 releases) and EEP Components and OS Support. |
Source on GitHub | https://github.com/mapr/hive/tree/2.3.3-mapr-1808 |
GitHub Release Tag |
2.3.3-mapr-1808 |
Maven Artifacts | See Maven Artifacts for the HPE Ezmeral Data Fabric. |
Package Names | Navigate to https://package.ezmeral.hpe.com/releases/MEP/, and select your EEP and OS to view the list of package names. |
Feature Support
- EEP 6.0.0 supports Hive-2.3.3 on Tez-0.9.
For more information, see Tez 0.9.1-1808 (EEP 6.0.0) Release Notes.
- EEP 6.0.0 does not support Hive on Spark, so you cannot use Spark as an execution engine for
Hive.
However, you can run Hive and Spark on the same cluster. You can also use Spark SQL and Drill to query Hive tables.
- EEP 6.0.0 does not support HDFS encryption in Hive tables.
- EEP 6.0.0 does not support HBase with Hive-2.3.3 starting from the 6.0.0 release.
- EEP 6.0.0 does not support LLAP with Hive-2.3.3, because Apache Slider is not a MapR ecosystem component.
- Hive 2.1 and later needs to run the
schematool
command as an initialization step.
New Features
MERGE
operator for HPE Ezmeral Data Fabric Database JSON Tables in Tez modeFor more information, see Understanding the MERGE Statement.
hive.metastore.allow.new.db.in.existing.directory
property with default value as falseYou must sethive.metastore.allow.new.db.in.existing.directory=true
in thehive-site.xml
file to create a database if the directory already exists and then restart Hive Metastore and HiveServer2.CAUTION You cannot set thehive.metastore.allow.new.db.in.existing.directory
using the Hive CLI or Beeline.- Ability to read default SSL passwords and keystore location from
/opt/mapr/conf/ssl-client.xml
Changes in Security with Default Configuration
-
The following properties are removed from the default
hive-site.xml
configuration on a secured cluster: -
Property Value hive.server2.webui.keystore.path
/opt/mapr/conf/ssl_keystore
hive.server2.webui.keystore.password
Default keystore password
- The following property is added to the
hive-site.xml
configuration by default on a secured cluster:Property Value hive.server2.use.SSL
true
- Because HiveServer2 is configured to use SSL encryption by default starting from
Hive-2.3 MEP-6.0.0, you must add
ssl=true;
to a JDBC connection string when using PAM or MapR-SASL authentication. - You can configure the JDBC connection string with SSL enabled or disabled.
For more information, refer to Configuring JDBC Connection String with SSL Encryption Enabled or Disabled.
- World-readable permissions enabled for Hive configuration files.
Known Issues
- In HIVE-19502, you cannot insert values into a table stored by
JdbcStorageHandler
. - You cannot connect to a Spark Thrift Server on a Kerberos-secured cluster, because Kerberos and SSL are not compatible.
However, you can modify the
hive.server2.use.SSL
tofalse
in thehive-site.xml
file. - In MAPR-HIVE-302, you cannot interact with Hive tables on PostgreSQL after upgrade.
- In Bug 32349, Simple Fetch from MapR Database JSON tables does not work in the Hive 2.3
release. Workaround: Set
hive.fetch.task.conversion=none
in thehive-site.xml
file or using the Hive CLI. Thehive.fetch.task.conversion
property is used for query conversion. Some select queries can be converted to a single FETCH task that minimizes the latency. Currently, the query should be single sourced and not have any subquery and also must not have any aggregations or distincts (which incurs RS - ReduceSinkOperator, requiring a MapReduce task), lateral views and joins:none
: disable hive.fetch.task.conversionminimal
: SELECT star, filter on partition columns, LIMIT onlymore
: SELECT, filter, LIMIT only (supports TABLESAMPLE and virtual columns)
- A non-administrative user can install hooks, which could represent a security vulnerability. See Preventing a Non-Administrative User from Installing Hooks for more information.
- Vectorized execution is a new Hive feature that can show performance improvements in some cases and cause stability issues with others. The Hive vectorized execution feature has many bugs in Hive 2.x. It is recommended to turn off this feature at a system level and only use it for certain queries which work fine using it. You must evaluate the benefit of this feature against the potential stability issues on a case by case basis.
Fixes
The Hive 2.3.3-1808 (EEP 6.0.0) release includes the following fixes on the base Apache release. For complete details, refer to the commit log for this project in GitHub.
Commit | Date (YYYY-MM-DD) | Comment |
---|---|---|
5ea2352 | 2018-08-08 | MAPR-HIVE-300: Hive HCat does not create a proper "client" on MapR-SASL cluster without Hive metastore |
c50c43b | 2018-07-24 | MSEN-11: You can save a file with a newline at the end |
d009cf0 | 2018-07-21 | MAPR-HIVE-281: WebHCat issues an error during an attempt to run a Hive query |
56ea17e | 2018-07-17 | MAPR-HIVE-285: The configure.sh fails when a custom security flag is set |
34180bf | 2018-07-09 | MAPR-HIVE-272: The HiveMapRDBJsonOutputFormat class must implement the HiveOutputFormat interface |
af94d85 | 2018-06-17 | MAPR-31641: The failed queries with invalid table alias or column reference are deleted |
9c572f7 | 2018-06-27 | MAPR-31803: The fix for Bug 30031 results in users not being able to create databases on an existing location |
1855442 | 2018-06-21 | MAPR-HIVE-242: The configure.sh script adds two rows to the end of hive-site.xml after every launching |
cf0655e | 2018-06-19 | MAPR-HIVE-258: The Zookeeper version in Hive is updated from 3.4.6 to 3.4.11 |
61773b5 | 2018-06-19 | MAPR-HIVE-244: The Hive configure.sh script overwrites properties by secure by default values and does not take into account the .not_configured_yet flag |
836f751 | 2018-06-19 | MAPR-HIVE-251: The default SSL password is read from the /opt/mapr/conf/ssl-server.xml |
28025b1 | 2018-06-22 | MAPR-HIVE-262: Only client passwords from ssl-client.xml are used |
3c463a5 | 2018-06-01 | MAPR-HIVE-223: NPE during CREATE ROLE uses SQL Standard-Based Hive Authorization |
0a458b4 | 2018-05-22 | MAPR-31380: The HeartBeat thread uses a cancelled delegation token while connecting to meta |
bc7a6bd | 2018-05-22 | MAPR-HIVE-212: SSL is setup by default when PAM plus MapR-SASL is configured by default |
39ae559 | 2018-05-12 | MAPR-HIVE-184: SSL encryption with PAM authentication is configured on the MapR-SASL secured cluster |
1f92945 | 2018-05-07 | MAPR-HIVE-208: Import to Hive as a parquet data format failed and could not initialize class org.apache.derby.jdbc.AutoloadedDriver40 |
f371524 | 2018-04-17 | MAPR-HIVE-193: The MERGE syntax for HPE Ezmeral Data Fabric Database JSON tables is implemented |
dfc7cfd | 2018-04-24 | MAPR-31175: The hive.exec.tmp.maprfsvolume should be false when operating with Tez |
19fd04f | 2018-04-13 | MAPR-HIVE-196: Running the configure.sh -R reconfigures the underlying MetastoreDB to derby in the hive-site.xml file |
5cfebda | 2018-04-12 | MAPR-27663: PidFilePatternConverter does not append the PID to the log name |
d842904 | 2018-04-11 | MAPR-HIVE-190: The log writes in two Hive.log files instead of one |
61735a1 | 2018-04-11 | MAPR-HIVE-171: A bash script for MapR configuration tool is added |
1f12b05 | 2018-04-11 | MAPR-HIVE-167: A sub-module API for configuring hadoop.security.credential.provider.path is added into the MapR configuration |
72c75e7 | 2018-04-11 | MAPR-HIVE-165: The HiveEncryption tool is moved to a separate sub-module and split into a CLI processing class and an Encryption processing class |
bd611a2 | 2018-04-10 | MAPR-HIVE-189: World-readable permissions for Hive configuration files are set |
efd951f | 2018-04-03 | MAPR-HIVE-187: The Hive CLI does not start on a 6.0.1 cluster with Spark 2.2.1 |
4f23f26 | 2018-03-21 | MAPR-30940: A Hive job fails by AccessControlException against files on the NM local disk |
c9445f0 | 2018-03-20 | MAPR-30895: Tez jobs are shown as KILLED in a RM UI |
The Hive 2.3.3-1808 (EEP 6.0.0) release also includes the following backported issues. For complete details, refer to the commit log for this project in GitHub.
Commit | Date | Comment |
---|---|---|
5b67110 | 2018-08-07 | HIVE-12408: SQLStdAuthorizer does not require an external table creator to be the owner of directory, in addition to read or write permissions |
e57f603 | 2018-07-25 | HIVE-17429: Hive JDBC does not return rows when querying Impala |
d1c113d | 2018-07-24 | HIVE-18393: Error returned when some other type is read as string from parquet tables |
224b05b | 2018-06-25 | HIVE-13000: Hive returns useless parsing error |
3b0918d | 2018-06-13 | HIVE-17963: Fix for HIVE-17113 is improved for non-blobstore file systems |
d313347 | 2018-06-12 | HIVE-17113: Duplicate bucket files are written to table by runaway task |
3eaba39 | 2018-05-31 | HIVE-17155: findConfFile() in HiveConf.java has issues with the configuration path |
800c6e5 | 2018-05-25 | HIVE-19649: Clean up inputs in JDBC PreparedStatement. Add unit tests |
24165d1 | 2018-04-05 | HIVE-18710: Extend inheritPerms to ACID in Hive 2.X |
Related Links
- Apache Hive 2.3.3 Release Notes
- Apache Hive 2.3.2 Release Notes
- Apache Hive 2.3.1 Release Notes
- Apache Hive 2.3.0 Release Notes