The following things should be considered when planning your upgrade to 3.7:
Your Fusion data sources from past releases are not automatically upgraded. You must manually edit them and reconfigure them after upgrading to this version. In addition, preexisting dashboards that used your Fusion data sources from previous versions will not work and will need to be recreated.
Due to the upgrade of the HashiCorp Consul (service discovery) version from 0.7.5 to 1.2.2, custom connectors built before Zoomdata 3.2 ( built for versions 3.1 or earlier versions) will not appear in Zoomdata after you migrate to Zoomdata 3.7. This is caused by incompatibilities between old connectors and the new Consul. To make your custom connectors compatible with Zoomdata 3.7 (and later versions), you must migrate them as described in Version 3.7 Custom Connector Migration Information.
This migration step is required only for custom connectors built with Zoomdata versions earlier than 3.2 or for connectors that were present before Zoomdata 3.2 but that are now deprecated (for example, the Hive On Tez connector, which was replaced by the Hive connector in Zoomdata 3.7).
If you have used Zoomdata versions earlier than 3.3, you must delete the RabbitMQ queues before you upgrade to Zoomdata 3.3 or later. In addition, an additional header specifying the media type on non-browser HTTP requests is now required. These changes are required because in Zoomdata version 3.3, we updated the Spring Cloud library underlying our application. See Version 3.3 Upgrade Information.
Zoomdata version 3.2 and later versions use headless Google Chrome instead of Firefox for Zoomdata's Screenshot service. The Chrome-based Screenshot service cannot be installed in CentOS v6 environments because the Google Chrome dependencies do not support that platform. See Setting Up the Screenshot Feature.
Was this topic helpful?