I don't mind maintaining minimal HBase support in Impala.
HBase related code does not seem to require a lot of maintenance.
That being said, it is probably OK to do all HBase tests and infrastructure
download only during exhaustive tests.

Regards,
Riza

On Wed, Feb 19, 2025 at 9:41 AM Xiang Yang <yx91...@126.com> wrote:

> Hi Michael, What I mean is that most users choose to use impala with kudu
> or iceberg, few users choose to use impala with HBase, and there are few
> HBase related patches in the community recently.
>
> Of course, this is my personal opinion, for reference only.
>
>
>
>
> At 2025-02-19 00:26:57,"Michael Smith" <michael.sm...@cloudera.com>,
> said:
> >I'm confused about what you mean "hasn't been active for a long time." Do
> >you mean HBase hasn't? They had releases last year.
> >
> >I'm not sure how I'd evaluate whether our support for HBase has been
> >active. I don't think fixes have been needed.
> >
> >On Thu, Feb 13, 2025 at 6:45 PM Xiang Yang <yx91...@126.com> wrote:
> >
> >> It's good to see this discussion, especially the deprecation of
> python2.By
> >> the way, I think It's also time to deprecate the support of HBase, which
> >> hasn't been active for a long time.
> >>
> >> Regards,
> >> Xiang Yang
> >>
> >>
> >> At 2025-02-11 04:55:20, "Laszlo Gaal" <lasz...@apache.org> wrote:
> >> >As mentioned earlier in
> >> >https://lists.apache.org/thread/96t7l66x8hlc5dswyt4g1mpn6wr5op4z, I
> would
> >> >like to start a discussion about Impala deprecating support for older
> >> >technologies that were retired by their creators:
> >> >- Python 2: Officially retired on Jan 1st, 2020 (
> >> >https://www.python.org/doc/sunset-python-2)
> >> >Operating system versions:
> >> >- CentOS 7: End of Maintenance reached on June 30th, 2024 (
> >> >
> >>
> https://blog.centos.org/2023/04/end-dates-are-coming-for-centos-stream-8-and-centos-linux-7/
> >> >, https://endoflife.date/centos)
> >> >- SLES 12: General support for 12.5 (the last minor release) ended on
> Oct
> >> >31st, 2024
> >> >- Ubuntu 16.04: Standard support ended in April 2021 (
> >> >https://ubuntu.com/about/release-cycle)
> >> >- Ubuntu 18.04: Standard support ended in April 2023
> >> >Client connectivity:
> >> >- support for the Beeswax client protocol: the default protocol for
> >> >impala-shell was changed to HS2 in August 2020
> >> >(commit 179b14876d13353302d075be13f0ed1145041bfa), almost five years
> ago.
> >> >This should have been long enough for all existing Impala clients to
> >> >embrace HS2 as the client protocol, so Beeswax can now be marked for
> >> >removal in the future.
> >> >
> >> >- Support for Java 8 is another interesting question. Major OpenJDK
> >> >rebuilds (Red Hat, Eclipse Temurin) show Java 8 support for 12-18
> months
> >> >moreIceberg removed support for Java 8 starting with their 1.7.0
> release,
> >> >so Impala will need to consider this any time we want to pick up any
> >> >Iceberg release higher than 1.6.x. Considering all this, I don't think
> we
> >> >need to make a decision on Java 8 support right now.
> >> >
> >> >Removing support for the above list certainly qualifies as a breaking
> >> >change, which would prompt a major version change.
> >> >The currently running Impala 4.5.0 release still contains support for
> all
> >> >of the above, so this release could be a good opportunity to branch
> Impala
> >> >4.x and start working on Impala v5.
> >> >
> >> >This is a bit like the way Impala 4.x was started with
> >> >https://lists.apache.org/thread/vkq8b4sb4j7x8yykbr92rw8c97k9n94q; so
> in a
> >> >similar way I propose the following list:
> >> >- release Impala 4.5.0 (the current release) with release notes
> expressly
> >> >deprecating an agreed-upon set of technologies,
> >> >- bump the version number to 5.0 on the master branch,
> >> >- open the gates for breaking changes
> >> >- eventually release v5.0 with breaking changes
> >> >
> >> >Creating a 4.x branch could be the subject of a separate (but related)
> >> >discussion.
> >> >
> >> >Please share your thoughts,
> >> >Thank you,
> >> >
> >> >  - LaszloG
> >>
>

Reply via email to