MR3

Here is a quick summary of performance comparison between MR3 1.9 and MR3 1.10 on the original TPC-DS

1. For the NPE error, we don't have test cases other that q tests included in Hive. If you know

Hello, If you use MR3 1.7 on Kubernetes, please update mr3-run-k8s scripts and switch to MR3 1.9. The

I updated the mr3-run-k8s repo to fix a couple of problems: 1) Java 17 options; 2) running superset

Hello, I have published a new blog article 'Performance Tuning for Single-table Queries'. It

Hive-MR3 is now shipped in HiDataPlus. For the latest announcement of HiDataPlus, please see: https:/

MR3 1.8 has been released. For changes, see the release notes: https://mr3docs.datamonad.com/docs/

Is there a chance that this slow S3 access is due to the use of Avro files? For example, do you see

Some users reported Hive-MR3 getting slower after running for a while. We reproduced this problem by

Thank you for sharing the story. I would like add that with the Tez runtime (in HDP and Hive-MR3),

We have uploaded MR3 1.8 release candidate. https://github.com/mr3project/mr3-release/releases/tag/v1

For those interested in Hive-MR3-Celeborn (with K8s) I know we had to bump the local storage on the

Continuing the comparison between Hive-Tez and Hive-MR3, here is the result of concurrent tests. We

Before the impending release of MR3 1.8, we would like to announce the release of Hive-MR3 with

Okay. I'll try to put together a list of things we've had to customize or do differnently. I

This will be fixed in MR3 1.8. --- Sungwoo On Wednesday, October 12, 2022 at 1:24:06 AM UTC+9

I think it's better to add slack entry on the home page of MR3. Sungwoo Park <glapark@gmail.

With the release of MR3 1.7, I published a blog article where we compare Trino, Spark, and Hive on

In MR3 1.7, we have changed the values for hive.auto.convert.join.noconditionaltask.size and tez.

MR3 1.7 has been released. I have updated the binary releases at Github and the Docker image at

Alright, sounds good. I suspect it is a problem with Hadoop, however, because the logic for

No. Currently MR3 can specify only a job queue (with the config key mr3.queue.name). It's easy to

On Hadoop, Java command is controlled by Hadoop, so running MR3 with its own Java 11 environments is

If you should use a different MySQL connector, please check out this page and the link to

Hello, Perhaps Spark 3.3.2 is faster than Hive 3 on your query. I think this is an issue with

Quick start guide for running Hive on MR3 on Kubernetes has been expanded. https://mr3docs.datamonad.

Great On Tuesday, January 31, 2023 at 10:14:59 AM UTC+8 da...@istwok.net wrote: Thanks for the

If Kyuubi can accelerate the execution of Spark workloads, you could ask the Kyuubi community for the

If we add UDF to “hive.aux.jars.path”, it will cause a serious problem: The JAR package added by the

Thanks On Thursday, December 15, 2022 at 7:48:58 PM UTC+8 Sungwoo Park wrote: You want to use hive.