mirror of https://github.com/apache/druid.git
e855c7fe1b
* Allow Cloud SegmentKillers to be instantiated without segment bucket or path This change fixes a bug that was introduced that causes ingestion to fail if data is ingested from one of the supported cloud storages (Azure, Google, S3), and the user is using another type of storage for deep storage. In this case the all segment killer implementations are instantiated. A change recently made forced a dependency between the supported cloud storage type SegmentKiller classes and the deep storage configuration for that storage type being set, which forced the deep storage bucket and prefix to be non-null. This caused a NullPointerException to be thrown when instantiating the SegmentKiller classes during ingestion. To fix this issue, the respective deep storage segment configs for the cloud storage types supported in druid are now allowed to have nullable bucket and prefix configurations * * Allow google deep storage bucket to be null |
||
---|---|---|
.. | ||
src | ||
pom.xml |