Cloudera Docs

Accessing and generating diagnostic bundles in Cloudera Data Warehouse on premises

Cloudera Data Warehouse collects diagnostic data on workload logs, such as Impala Coordinator, Statefulset, CatalogD logs and stores it in the tmp directory on HDFS. You can download the logs using the Hue File Browser from the base cluster.

During the lifetime of a cluster, logs are continously written to the following directory on HDFS: [**WAREHOUSE-DIR**]/warehouse/tablespace/external/hive/sys.db/. When you click Collect Diagnostic Bundle from the Cloudera Data Warehouse web interface, Cloudera Data Warehouse collects the logs for the specified time interval and for the services that you select. These logs are compressed in a ZIP file format and stored in the tmp directory.

  1. To check the job status and find the HDFS location where the logs are stored, select Edit from the Virtual Warehouse options menu and go to the DIAGNOSTIC BUNDLE tab.
    The logs are collected and bundled under the /tmp/[***VIRTUAL-WAREHOUSE-ID-TIMESTAMP***].zip directory.
  2. To access and download the logs, open the Hue service from the base cluster.
  3. Go to the Hue File Browser and click the forward slash (/) before the user directory as shown in the following image:

    Screenshot showing how to go to the tmp directory to download the diagnostic bundles from the Hue File Browser.
    The tmp directory is displayed. You can access and download the logs to your computer by clicking Download.