[
https://issues.apache.org/jira/browse/HIVE-24485?focusedWorklogId=520248&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-520248
]
ASF GitHub Bot logged work on HIVE-24485:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 04/Dec/20 16:18
Start Date: 04/Dec/20 16:18
Worklog Time Spent: 10m
Work Description: okumin opened a new pull request #1744:
URL: https://github.com/apache/hive/pull/1744
### What changes were proposed in this pull request?
Make it possible to apply `tez.shuffle-vertex-manager.min-src-fraction` and
`tez.shuffle-vertex-manager.max-src-fraction` even if auto reducer parallelism
is enabled.
https://issues.apache.org/jira/browse/HIVE-24485
### Why are the changes needed?
With this PR, we can tweak the trade-off between timing to start and
accuracy of estimation.
Tez can gather more samples with higher fractions while it delays the start
of the next vertex.
### Does this PR introduce _any_ user-facing change?
Users who are configuring
`tez.shuffle-vertex-manager.{min,max}-src-fraction` or
`mapreduce.job.reduce.slowstart.completedmaps`, the alternative of
`tez.shuffle-vertex-manager.min-src-fraction`, can face the change of behavior
when they enable auto reducer parallelism.
https://github.com/apache/tez/blob/dadc09f5a44c1cb61af00efecb3d27b92c92aa8f/tez-mapreduce/src/main/java/org/apache/tez/mapreduce/hadoop/DeprecatedKeys.java#L123
### How was this patch tested?
No test cases will be added because `Vertex#VertexManagerPluginDescriptor`
is invisible from the outside of a Tez package.
We can check the change by running a job and then checking the Tez log.
```
beeline -e '
SET hive.tez.auto.reducer.parallelism=true;
SET hive.tez.min.partition.factor=1.0; -- enforce auto-parallelism
SET tez.shuffle-vertex-manager.min-src-fraction=0.55;
SET tez.shuffle-vertex-manager.max-src-fraction=0.95;
CREATE TABLE mofu (name string);
INSERT INTO mofu (name) VALUES ('12345');
SELECT name, count(*) FROM mofu GROUP BY name;
'
```
```
2020-12-04 16:10:47,170 [INFO] [Dispatcher thread {Central}]
|vertexmanager.ShuffleVertexManagerBase|: Settings minFrac: 0.55 maxFrac: 0.95
auto: true desiredTaskIput: 256000000
```
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
Issue Time Tracking
-------------------
Worklog Id: (was: 520248)
Remaining Estimate: 0h
Time Spent: 10m
> Make the slow-start behavior tunable
> ------------------------------------
>
> Key: HIVE-24485
> URL: https://issues.apache.org/jira/browse/HIVE-24485
> Project: Hive
> Issue Type: Improvement
> Components: Hive, Tez
> Affects Versions: 3.1.2, 4.0.0
> Reporter: okumin
> Assignee: okumin
> Priority: Major
> Time Spent: 10m
> Remaining Estimate: 0h
>
> This ticket would enable users to configure the timing of slow-start with
> `tez.shuffle-vertex-manager.min-src-fraction` and
> `tez.shuffle-vertex-manager.max-src-fraction`.
> Hive on Tez currently doesn't honor these parameters and ShuffleVertexManager
> always uses the default value.
> We can control the timing to start vertexes the accuracy of estimated input
> size if we can tweak these ones. This is useful when a vertex has tasks that
> process a different amount of data.
>
> We can reproduce the issue with this query.
> {code:java}
> SET hive.tez.auto.reducer.parallelism=true;
> SET hive.tez.min.partition.factor=1.0; -- enforce auto-parallelism
> SET tez.shuffle-vertex-manager.min-src-fraction=0.55;
> SET tez.shuffle-vertex-manager.max-src-fraction=0.95;
> CREATE TABLE mofu (name string);
> INSERT INTO mofu (name) VALUES ('12345');
> SELECT name, count(*) FROM mofu GROUP BY name;{code}
> The fractions are ignored.
> {code:java}
> 2020-12-04 11:41:42,484 [INFO] [Dispatcher thread {Central}]
> |vertexmanager.ShuffleVertexManagerBase|: Settings minFrac: 0.25 maxFrac:
> 0.75 auto: true desiredTaskIput: 256000000
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
