Hi,
We deployed a test version of Hive 4 in our production. It is mostly
working well. I appreciate the community's effort.
Let me bump this thread so that community members can share the latest
situation, and we can contribute to the remaining problems.
Currently, we list the following tickets labeled with hive-4.0.0-must[1].
- HIVE-27858: OOM happens when selecting many columns and JOIN
- HIVE-26654: Test with the TPC-DS benchmark
- HIVE-24167: No.14 (Postponed until 4.1.0)
- HIVE-26986: No.71 (No labels)
We have already applied a workaround[2] for HIVE-24167. Also, Denys
says HIVE-26986 is not a blocker of Hive 4.0.0. So, I expect
HIVE-27858 to be the only blocker.
I have some proposals about the next action.
- Can we take over HIVE-27858? We are the reporter of the ticket. It
would be reasonable If Naveen is not working on it yet
- We may review HIVE-27984[3]. As I mentioned in another thread[4], it
would have a fairly big impact on the Hive ecosystem
And I would like someone to take a look at the following PR for
HIVE-24167. It is not a blocker thanks to
`hive.optimize.cte.materialize.threshold=-1`. However, we plan to use
Hive 4 with CTE materialization.
https://github.com/apache/hive/pull/5037
Thanks!
- [1] https://issues.apache.org/jira/issues/?jql=labels%20%3D%20hive-4.0.0-must
- [2] https://github.com/apache/hive/pull/4858
- [3] https://github.com/apache/hive/pull/4984
- [4] https://www.mail-archive.com/dev@hive.apache.org/msg144280.html
Regards,
Okumin
On Thu, Nov 23, 2023 at 11:42 PM Butao Zhang <butaozha...@163.com> wrote:
>
> Cool! Thank you Denys for driving this!
>
>
> Thanks,
> Butao Zhang
> ---- Replied Message ----
> | From | Attila Turoczy<aturo...@cloudera.com.INVALID> |
> | Date | 11/23/2023 22:29 |
> | To | <dev@hive.apache.org> |
> | Subject | Re: Release of Hive 4 and TPC-DS benchmark |
> Excellent new Denys! Hive 4 is here! Can't wait :)
>
> -Attila
>
>
> On Thu, Nov 23, 2023 at 3:20 PM Denys Kuzmenko <dkuzme...@apache.org> wrote:
>
> Update:
> 1. Query 2, 71: Resolved in HIVE-27006 [1];
>
> 2. Query 97: Under review in HIVE-27269 [2];
> Thanks, Seonggon for providing a reproduce qfile.
>
> 3. Query 14: Reported in HIVE-24167 [3];
> set hive.optimize.cte.materialize.threshold to -1 by default in Hive 4 and
> fix it in the next versions
>
> 4. HIVE-26986 [4] is a performance improvement that is nice to have, but
> not a blocker for the release.
>
> Be advised, next week we plan to cut 4.0.0 release branch from master and
> start testing.
>
> Thanks, Denys
>
> [1] https://issues.apache.org/jira/browse/HIVE-27006
> [2] https://issues.apache.org/jira/browse/HIVE-27269
> [3] https://issues.apache.org/jira/browse/HIVE-24167
> [4] https://issues.apache.org/jira/browse/HIVE-26986
>
>
