Rolling Upgrade Guide
Also available as:
PDF
loading table of contents...

Downgrade Knox

At this stage, the following components have been upgraded and need to be downgraded:

  • Knox

  • Non-core cluster components: Falcon, Oozie, Hive

  • Core slave nodes: DataNodes, NodeManager, RegionServer

  • Core masters: HDFS, YARN, HBase

  • Ranger

  • ZooKeeper

There is service disruption when downgrading Knox. Downgrading Knox requires bringing Knox down completely, downgrading, and starting up again.

  1. Shut down all Knox servers:

    su -l knox "/usr/hdp/2.2.9.0-$BUILD/knox/bin/gateway.sh stop"

  2. Revert any changes to the data that were made after the upgrade process started, or restore the earlier version of your data from backup.

  3. Switch to the previous Knox software version, and then start the server:

    hdp-select set knox-server 2.2.0.0-2041

    su -l knox "/usr/hdp/2.2.0.0-2041/knox/bin/gateway.sh start"

  4. Validate the installation:

    1. Check /var/log/knox/gateway.log for errors or successful start.

    2. Validate cluster access through a WebHDFS API such as LISTSTATUS:

      curl -ivk -u {user}:{password}

      https://{knox-host}:8443/gateway/webhdfs/v1/tmp?op=LISTSTATUS

    3. If the Knox admin service is deployed to the gateway instance, the version API should indicate the previous version of the Knox binaries:

      curl -ivk -u {adminuser}:{adminpassword}

      https://{knox-host}:8443/gateway/admin/v1/version

Continue with the next downgrade step.