hawq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Radar Lei <r...@pivotal.io>
Subject Re: Re: Re: Revisit plan for Apache HAWQ 2.4.0.0 release with Hive access support in pluggable storage framework
Date Wed, 01 Aug 2018 08:00:23 GMT
Now seems all the issues in 2.4.0.0-incubating scope are addressed. I plan
to start the release process this week. Thanks all!

Regards,
Radar

On Mon, Jul 9, 2018 at 4:08 PM, Ruilong Huo <huor@apache.org> wrote:

> Sounds a good idea. Thanks Radar for taking care of that!
>
>
> Best regards,
> Ruilong Huo
> At 2018-07-09 16:02:54, "Radar Lei" <rlei@pivotal.io> wrote:
> >Hi Ruilong,
> >
> >As there are no further comments on this topic, I will remove the
> >feature 'Support
> >accessing Hive table data by Pluggable Storage Framework' from current
> >Apache HAWQ 2.4.0.0-incubating release.
> >
> >The updated features including in current release are listed as:
> >
> >   - New Feature: Pluggable Vectorized Execution Engine on HAWQ.
> >   - New Feature: Support Runtime Filter for HAWQ local hash join.
> >
> >
> >Since the two list features are all completed, I will start the release
> >process soon.
> >
> >Thanks.
> >
> >
> >Regards,
> >Radar
> >
> >On Tue, Jul 3, 2018 at 9:18 AM, HuoRuilong <huoruilong@163.com> wrote:
> >
> >> Exactly, Ed. The plan is to split the work and make them available in
> >> subsequent releases. For now, there are five sub tasks tracking them
> >> (HAWQ-1628 ~ HAWQ-1632). Though we might separate them even more as we
> go
> >> deeper with it.
> >>
> >> In fact, the work for the initial task has been started. We might
> discuss
> >> which releases these features will fit into as we drafting the plan for
> >> future releases.
> >>
> >>
> >> Best regards,
> >> Ruilong
> >>
> >>
> >> At 2018-07-02 22:27:52, "Ed Espino" <espino@apache.org> wrote:
> >> >Thanks for bring this up Ruilong. Given your input, It is reasonable to
> >> >remove the pluggable storage framework features from 2.4.0.0. To keep
> the
> >> >momentum going, going forward, can we separate the work and
> incrementally
> >> >commit prerequisite work over several releases?
> >> >
> >> >-=e
> >> >
> >> >On Mon, Jul 2, 2018 at 2:34 AM Radar Lei <rlei@pivotal.io> wrote:
> >> >
> >> >> Hi Ruilong,
> >> >>
> >> >> Thanks for the update, for me, I'm fine with removing this feature
> from
> >> >> 2.4.0.0 release since the scope is bigger than expected.
> >> >>
> >> >> And we still have Vectorized Execution and Runtime Filter features
in
> >> this
> >> >> release, they are good enough to make a new release.
> >> >>
> >> >> Regards,
> >> >> Radar
> >> >>
> >> >> On Mon, Jul 2, 2018 at 4:03 PM, Ruilong Huo <huor@apache.org>
wrote:
> >> >>
> >> >> > Hi Radar and All,
> >> >> >
> >> >> >
> >> >> > Had a closer look at the plan for Apache HAWQ 2.4.0.0 release,
it
> >> turns
> >> >> > out that the support of Hive data access in pluggable storage
> >> framework
> >> >> has
> >> >> > larger content than expected. In fact, to make Hive protocol
> >> available,
> >> >> the
> >> >> > prerequisite need to be done at first, which including HDFS
> protocol,
> >> >> ORC,
> >> >> > TEXT/CSV format, etc. The corresponding task have been added in
> >> umbrella
> >> >> > feature HAWQ-786 as below:
> >> >> >
> >> >> >
> >> >> > HAWQ-1628:  Support HDFS protocol using pluggable storage framework
> >> >> > HAWQ-1629:  Support ORC format using pluggable storage framework
> >> >> > HAWQ-1630:  Support TEXT/CSV format using pluggable storage
> framework
> >> >> > HAWQ-1631:  Support Hive protocol using pluggable storage framework
> >> >> > HAWQ-1632:  Wrap up pluggable storage framework and HDFS/Hive
> >> >> ORC/TEXT/CSV
> >> >> > support
> >> >> >
> >> >> >
> >> >> > Given that, I think it will be better if we can release 2.4.0.0
> >> without
> >> >> > this feature for now and accommodate it to subsequent releases.
Any
> >> >> comment?
> >> >> >
> >> >> >
> >> >> > Best regards,
> >> >> > Ruilong Huo
> >> >> >
> >> >> >
> >> >> >
> >> >> >
> >> >> >
> >> >> >
> >> >>
> >>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message