[
https://issues.apache.org/jira/browse/FLINK-13238?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Leonard Xu updated FLINK-13238:
-------------------------------
Description:
The blink planner has an independent CI profile, but still exceeded 50 min
limit from time to time. We need to optimize the tests to reduce the testing
time. This will leading to Travis failure.
We need do some work to reduce time:
1.Optimizer big tests:
192.503 s LongHashTableTest
83.969 s BinaryExternalSorterTest
261.497 s BinaryHashTableTest
74.223 s - in org.apache.flink.table.runtime.stream.sql.RankITCase
135.375 s - in org.apache.flink.table.runtime.stream.sql.JoinITCase
99.007 s - in org.apache.flink.table.runtime.stream.sql.SplitAggregateITCase
61.216 s - in org.apache.flink.table.runtime.stream.sql.OverWindowITCase
77.409 s - in
org.apache.flink.table.runtime.stream.sql.SemiAntiJoinStreamITCase
83.83 s - in org.apache.flink.table.runtime.stream.sql.AggregateRemoveITCase
314.376 s - in org.apache.flink.table.runtime.stream.sql.AggregateITCase
121.19 s - in org.apache.flink.table.runtime.stream.table.JoinITCase
74.417 s - in
org.apache.flink.table.runtime.batch.sql.agg.SortDistinctAggregateITCase
109.185 s - in org.apache.flink.table.runtime.batch.sql.agg.HashAggITCase
178.181 s - in
org.apache.flink.table.runtime.batch.sql.agg.AggregateReduceGroupingITCase
112.006 s - in org.apache.flink.table.runtime.batch.sql.agg.SortAggITCase
61.863 s - in org.apache.flink.table.runtime.batch.sql.agg.GroupingSetsITCase
62.941 s - in
org.apache.flink.table.runtime.batch.sql.agg.HashDistinctAggregateITCase
64.58 s - in org.apache.flink.table.runtime.batch.sql.CalcITCase
81.272 s - in org.apache.flink.table.runtime.batch.sql.OverWindowITCase
75.298 s - in org.apache.flink.table.runtime.batch.sql.join.JoinITCase
82.923 s - in org.apache.flink.table.runtime.batch.sql.join.OuterJoinITCase
145.538 s - in org.apache.flink.table.runtime.batch.sql.join.SemiJoinITCase
214.933 s - in org.apache.flink.table.runtime.batch.sql.join.InnerJoinITCase
2.Reuse miniCluster in ITCases.
Every MiniCluster initialization takes 15 seconds, and MiniCluster is
class-level reuse. We have many ITCase classes.
was:
The blink planner has an independent CI profile, but still exceeded 50 min
limit from time to time. We need to optimize the tests to reduce the testing
time. This will leading to Travis failure.
We need do some work to reduce time:
1.Optimizer big tests:
192.503 s LongHashTableTest
83.969 s BinaryExternalSorterTest
261.497 s BinaryHashTableTest
74.223 s - in org.apache.flink.table.runtime.stream.sql.RankITCase
135.375 s - in org.apache.flink.table.runtime.stream.sql.JoinITCase
99.007 s - in org.apache.flink.table.runtime.stream.sql.SplitAggregateITCase
61.216 s - in org.apache.flink.table.runtime.stream.sql.OverWindowITCase
77.409 s - in org.apache.flink.table.runtime.stream.sql.SemiAntiJoinStreamITCase
83.83 s - in org.apache.flink.table.runtime.stream.sql.AggregateRemoveITCase
314.376 s - in org.apache.flink.table.runtime.stream.sql.AggregateITCase
121.19 s - in org.apache.flink.table.runtime.stream.table.JoinITCase
74.417 s - in
org.apache.flink.table.runtime.batch.sql.agg.SortDistinctAggregateITCase
109.185 s - in org.apache.flink.table.runtime.batch.sql.agg.HashAggITCase
178.181 s - in
org.apache.flink.table.runtime.batch.sql.agg.AggregateReduceGroupingITCase
112.006 s - in org.apache.flink.table.runtime.batch.sql.agg.SortAggITCase
61.863 s - in org.apache.flink.table.runtime.batch.sql.agg.GroupingSetsITCase
62.941 s - in
org.apache.flink.table.runtime.batch.sql.agg.HashDistinctAggregateITCase
64.58 s - in org.apache.flink.table.runtime.batch.sql.CalcITCase
81.272 s - in org.apache.flink.table.runtime.batch.sql.OverWindowITCase
75.298 s - in org.apache.flink.table.runtime.batch.sql.join.JoinITCase
82.923 s - in org.apache.flink.table.runtime.batch.sql.join.OuterJoinITCase
145.538 s - in org.apache.flink.table.runtime.batch.sql.join.SemiJoinITCase
214.933 s - in org.apache.flink.table.runtime.batch.sql.join.InnerJoinITCase
2.Reuse miniCluster in ITCases.
Every MiniCluster initialization takes 15 seconds, and MiniCluster is
class-level reuse. We have many ITCase classes.
> Reduce blink planner's testing time
> -----------------------------------
>
> Key: FLINK-13238
> URL: https://issues.apache.org/jira/browse/FLINK-13238
> Project: Flink
> Issue Type: Improvement
> Components: Table SQL / Planner
> Reporter: Kurt Young
> Assignee: Jingsong Lee
> Priority: Major
>
> The blink planner has an independent CI profile, but still exceeded 50 min
> limit from time to time. We need to optimize the tests to reduce the testing
> time. This will leading to Travis failure.
>
> We need do some work to reduce time:
> 1.Optimizer big tests:
> 192.503 s LongHashTableTest
> 83.969 s BinaryExternalSorterTest
> 261.497 s BinaryHashTableTest
> 74.223 s - in org.apache.flink.table.runtime.stream.sql.RankITCase
> 135.375 s - in org.apache.flink.table.runtime.stream.sql.JoinITCase
> 99.007 s - in org.apache.flink.table.runtime.stream.sql.SplitAggregateITCase
> 61.216 s - in org.apache.flink.table.runtime.stream.sql.OverWindowITCase
> 77.409 s - in
> org.apache.flink.table.runtime.stream.sql.SemiAntiJoinStreamITCase
> 83.83 s - in org.apache.flink.table.runtime.stream.sql.AggregateRemoveITCase
> 314.376 s - in org.apache.flink.table.runtime.stream.sql.AggregateITCase
> 121.19 s - in org.apache.flink.table.runtime.stream.table.JoinITCase
> 74.417 s - in
> org.apache.flink.table.runtime.batch.sql.agg.SortDistinctAggregateITCase
> 109.185 s - in org.apache.flink.table.runtime.batch.sql.agg.HashAggITCase
> 178.181 s - in
> org.apache.flink.table.runtime.batch.sql.agg.AggregateReduceGroupingITCase
> 112.006 s - in org.apache.flink.table.runtime.batch.sql.agg.SortAggITCase
> 61.863 s - in org.apache.flink.table.runtime.batch.sql.agg.GroupingSetsITCase
> 62.941 s - in
> org.apache.flink.table.runtime.batch.sql.agg.HashDistinctAggregateITCase
> 64.58 s - in org.apache.flink.table.runtime.batch.sql.CalcITCase
> 81.272 s - in org.apache.flink.table.runtime.batch.sql.OverWindowITCase
> 75.298 s - in org.apache.flink.table.runtime.batch.sql.join.JoinITCase
> 82.923 s - in org.apache.flink.table.runtime.batch.sql.join.OuterJoinITCase
> 145.538 s - in org.apache.flink.table.runtime.batch.sql.join.SemiJoinITCase
> 214.933 s - in org.apache.flink.table.runtime.batch.sql.join.InnerJoinITCase
>
> 2.Reuse miniCluster in ITCases.
> Every MiniCluster initialization takes 15 seconds, and MiniCluster is
> class-level reuse. We have many ITCase classes.
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
