Details
-
Improvement
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
3.0.0
-
None
Description
When a shuffle block (content) is fetched the network is always used even when it is fetched from the external shuffle service running on the same host. This can be avoided by getting the local directories of the same host executors from the external shuffle service and accessing those blocks from the disk directly.
Attachments
Issue Links
- causes
-
SPARK-30604 HostLocal Block size missed in log total bytes
- Resolved
- depends upon
-
SPARK-27622 Avoid the network when block manager fetches disk persisted RDD blocks from the same host
- Resolved
- is related to
-
SPARK-31055 Update config docs for shuffle local host reads to have dep on external shuffle service
- Resolved
- relates to
-
SPARK-6521 Bypass network shuffle read if both endpoints are local
- Resolved
-
SPARK-32077 Support host-local shuffle data reading with external shuffle service disabled
- Resolved
- links to