Known limitations in Hue
Review the known limitations in Hue.
Hue has the following limitations:
- Node depth for graphing Oozie workflows because of performance issues. See Improved Oozie Workflow display of large Graphs.
- You must use the Cloudera-provided Apache Load balancer to serve static content,
because:
- It serves static JavaScript, CSS, and Webpack files for client requests and reduces the load from the backend Python web server.
- The Hue load balancer uses a sticky cookie session to route requests to the same backend as the Python web server, which talks to the same coordinator.
- Hue can only show logs from either Spark1 or Spark2, not both at a time.
- Spark notebook is not supported.
- External RDBMS in the query editor is not supported out of the box by default. Cloudera support will assist on a best-effort basis. Cloudera recommends that you raise issues in the open-source github community.
- Impala queries stay in the “executing” state so that Hue can display results when users are ready
- We need to limit the amount of data available to download from Hive/Impala because massive downloads cause performance degradation. Multiple simultaneous downloads of result sets could also degrade performance.
- Upstream features and connectors may not function properly in CDP. Cloudera recommends that you raise issues in the open-source github community.