Skip to end of metadata
Go to start of metadata

This page is prepared for Hive committers. You need committer rights to create a new Hive release.


  1. Bulk update Jira to unassign from this release all issues that are open non-blockers and send follow-up notification to the developer list that this was done.
  2. Run 'ant rat' to and examine the generated report for any files, especially .java files should be having Apache license headers. Add the license header to any file that is missing it (open a jira and submit a patch).
  3. Update copyright date in NOTICE. If the components mentioned in them have updated version, you would need to update the copyright dates for those. (Thejas comment: It sounds like entries are needed in NOTICE only if the license requires such attribution . See -


Skip this section if this is NOT the first release in a series (i.e., release X.Y.0).

  1. Notify developers on the #hive IRC channel and dev@hive mailing lists that you are about to branch a release.

  2. Create a branch for the release series: 

  3. Increment the value of the version property in all pom.xml files. For example, if the current value is 0.7.0-SNAPSHOT, the new value should be 0.8.0-SNAPSHOT. Please note that the SNAPSHOT suffix is required in order to indicate that this is an unreleased development branch. This can be accomplished with a single command using Maven's Versions plugin as follows: 

  4. Make changes to metastore upgrade scripts. See HIVE-6555 on how this was done for HIVE 0.13.
  5. Verify that the build is working with changes.
  6. Commit these changes to trunk.

Updating Release Branch

These operations take place in the release branch.

  1. Check out the release branch with:

  2. Update the version property value in all pom.xml files. You should remove the SNAPSHOT suffix and set version equal to hive-X.Y.Z where Z is the point release number in this release series (0 for the first one, in which case this step is a no-op since you already did this above when creating the branch). Use Maven's Versions plugin to do this as follows:

  3. Update the the value of the TRACKING_BRANCH field in the .reviewboardrc file to point to the origin/branch-X.Y.
  4. Verify that the build is working with changes.
  5. Commit these changes.

  6. If not already done, merge desired patches from trunk into the branch and commit these changes. We don't currently use svn merge for this, but if we did, the script would be

  7. You probably also want to commit a patch (on both trunk and branch) which updates README.txt to bring it up to date (at a minimum, search+replacing references to the version number). Also check NOTICE to see if anything needs to be updated for recent library dependency changes or additions.
    1. Select all of the JIRAs for the current release that aren't FIXED and do bulk update to clear the 'Fixed Version' field.
    2. Likewise, use JIRA's Release Notes link to generate content for the RELEASE_NOTES.txt file. Be sure to select 'Text' format. (It's OK to do this with a direct commit rather than a patch.)
    3. Update the release notes in trunk with the release notes in branch.
  8. Tag the release candidate (R is the release candidate number, and also starts from 0):


  1. Make sure your release notes have been updated for any new commits, and go through the previous steps if necessary.
  2. Build the release (binary and source versions) after running unit tests. Manually create the md5 files. On a Mac use md5 in place of md5sum.

    % mvn install -Pdist -DskipTests -Phadoop-1  -Dmaven.javadoc.skip=true -DcreateChecksum=true
    % cd packaging/target
    % md5 apache-hive-0.13.0-bin.tar.gz > apache-hive-X.Y.Z-bin.tar.gz.md5
    % md5 apache-hive-0.13.0-src.tar.gz > apache-hive-X.Y.Z-src.tar.gz.md5
  3. Verify that the MD5 checksums are valid:

    % md5sum -c hive-X.Y.Z.tar.gz.md5
    hive-X.Y.Z.tar.gz: OK
    % md5sum -c hive-X.Y.Z-bin.tar.gz.md5
    hive-X.Y.Z-bin.tar.gz: OK
  4. Check that release file looks ok -- e.g., install it and run examples from tutorial.
  5. Setup your PGP keys for signing the release, if you don't have them already.
    1. See,
  6. Sign the release (see Step-By-Step Guide to Mirroring Releases for more information).

    % gpg --armor --output apache-hive-X.Y.Z-bin.tar.gz.asc --detach-sig apache-hive-X.Y.Z-bin.tar.gz
    % gpg --armor --output apache-hive-X.Y.Z-src.tar.gz.asc --detach-sig apache-hive-X.Y.Z-src.tar.gz
  7. Copy release files to a public place.

    % ssh mkdir public_html/hive-X.Y.Z-candidate-0
    % scp -p apache-hive-X.Y.Z*.tar.gz*
  8. Publish Maven artifacts to the Apache staging repository:

    % mvn deploy -DskipTests -Phadoop-1 -Papache-release -Dmaven.javadoc.skip=true
  9. Login to the Apache Nexus server and "close" the staged repository. This makes the artifacts available at a temporary URL.


  1. Call a release vote on dev at

    Subject: [VOTE] Apache Hive X.Y.Z Release Candidate N
    Apache Hive X.Y.Z Release Candidate N is available here:
    Maven artifacts are available here:
    Source tag for RCN is at:
    Voting will conclude in 72 hours.
    Hive PMC Members: Please test and vote.

Verifying the Release Candidate

  1. Verifying the PGP signature:

    #get the hive committers keys file
    gpg --import <keys file>
    gpg --verify hive-X.Y.Z-bin.tar.gz.asc  hive-X.Y.Z-bin.tar.gz
    gpg --verify hive-X.Y.Z.tar.gz.asc  hive-X.Y.Z.tar.gz
  2. Verifying the md5 checksum:
    See the step under Building.


Once three PMC members have voted for a release, it may be published.

  1. Tag the release:

    svn move \ -m "HiveX.Y.Z release."
  2. Copy release files to the distribution directory and make them writable by the hive group.

    cp -pR public_html/hive-X.Y.Z-candidate-0/ /www/
    cd /www/
    chgrp -R hive hive-X.Y.Z
    chmod -R g+w hive-X.Y.Z
  3. The release directory usually contains just two releases, the most recent from two branches, with a link named 'stable' to the most recent recommended version.

    cd /www/
    rm -rf hive-A.B.C; rm stable
    ln -s hive-A.B.D stable
  4. Wait 24 hours for release to propagate to mirrors.
  5. Prepare to edit the website.

    svn co
  6. Copy the new release docs to publish/docs/rX.Y.Z:

  7. If this is a bugfix release, delete the documentation for the obsoleted release.

  8. Update the documentation links in author/src/documentation/content/xdocs/site.xml.
  9. Update the release news in author/src/documentation/content/xdocs/releases.xml.
  10. Update publish/.htaccess with the current release and any new redirects for obsoleted releases.
  11. Regenerate the site, review it, then commit it. The live site is automatically deployed from subversion after a few minutes.

  12. Update JIRA
    1. Ensure that only issues in the "Fixed" state have a "Fix Version" set to release X.Y.Z.
    2. Release the version. Visit the "Administer Project" page, then the "Manage versions" page. You need to have the "Admin" role in Hive's Jira for this step and the next.
    3. Close issues resolved in the release. Disable mail notifications for this bulk change.
  13. Login to the Apache Nexus server and mark the release candidate artifacts as released.
  14. Send a release announcement to Hive user and dev lists as well as the Apache announce list. This email should be sent from your Apache email address:

    Subject: [ANNOUNCE] Apache Hive X.Y.Z Released
    The Apache Hive team is proud to announce the the release of Apache
    Hive version X.Y.Z.
    The Apache Hive (TM) data warehouse software facilitates querying and
    managing large datasets residing in distributed storage. Built on top
    of Apache Hadoop (TM), it provides:
    * Tools to enable easy data extract/transform/load (ETL)
    * A mechanism to impose structure on a variety of data formats
    * Access to files stored either directly in Apache HDFS (TM) or in other
      data storage systems such as Apache HBase (TM)
    * Query execution via Apache Hadoop MapReduce and Apache Tez frameworks.
    For Hive release details and downloads, please visit:
    Hive X.Y.Z Release Notes are available here: [UPDATE THIS LINK] 
    We would like to thank the many contributors who made this release
    The Apache Hive Team

See Also

  • No labels