Github user jiadexin commented on the issue:
https://github.com/apache/incubator-hawq/pull/972
@shivzoneIf there are no partitions, then only one fragment.
if have partitions ,HAWQ-MASTER scheduling strategy support is required.
HDFS is distributed, and pxf-hdfs can allocate hosts for the fragment via
the hdfs file metadata.
The role of pxf-jdbc is mainly used to integrate the enterprise has been
running the traditional relational database, these systems are generally
stand-alone system, may not deploy PXF Engine.
In the current HAWQ PXF engine implementation, pxf-instance can not know
all pxf-hosts, so only the current host name can be assigned.
if configured in the LOCATION in the DDL, when a large number of 'pxf
hosts' is used, the DDL statement is extremely long. and not flexible.
i think, HAWQ-MASTER wants to support a scheduling strategy: when the
fragment host name is null, HAWQ-MASTER automatically assigned a HAWQ-segment
host.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
