— Hive sink support
Flume does not provide a native sink that stores the data that can be directly consumed by Hive.
— HBase sink does not work out of the box
The HBase sink does not work out of the box in Flume; it fails to connect to Zookeeper.
- Remove the file /etc/zookeeper/conf/zoo.cfg from the Flume client machines, and specify the Zookeeper details in hbase-site.xml or flume.conf; OR
- Use CDH 4.5 or later with Cloudera Manager 4.7
or later. Cloudera Manager passes a flag that
causes Flume not to include any files named
zoo.cfg in its classpath at
startup time. The flag is enabled by default.Note
If you need to do so for backward compatibility purposes, you can disable this flag as follows:
- Go to Flume service -> Configuration -> Agent -> Advanced.
- Disable the HBase sink prefer hbase-site.xml over Zookeeper config property.
- Restart the Flume service.
— Fast Replay does not work with encrypted File Channel
If an encrypted file channel is set to use fast replay, the replay will fail and the channel will fail to start.
Bug: FLUME-1885 (unresolved as of 5/19/14)
Workaround: Disable fast replay for the encrypted channel by setting use-fast-replay to false.