[ 
https://issues.apache.org/jira/browse/FLINK-35606?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17861328#comment-17861328
 ] 

Rui Fan commented on FLINK-35606:
---------------------------------

Created https://issues.apache.org/jira/browse/FLINK-35738 to verify FLINK-26050

> Release Testing Instructions: Verify FLINK-26050 Too many small sst files in 
> rocksdb state backend when using time window created in ascending order
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: FLINK-35606
>                 URL: https://issues.apache.org/jira/browse/FLINK-35606
>             Project: Flink
>          Issue Type: Sub-task
>          Components: Runtime / State Backends
>            Reporter: Rui Fan
>            Assignee: Roman Khachatryan
>            Priority: Blocker
>              Labels: release-testing
>             Fix For: 1.20.0
>
>
> Follow up the test for https://issues.apache.org/jira/browse/FLINK-26050
> The problem occurs when using RocksDB and specific queries/jobs (please see 
> the ticket for the detailed description).
> To test the solution, run the following query with RocksDB as a state backend:
>  
> {code:java}
> INSERT INTO top_5_highest_view_time
> SELECT *
> FROM   (
>                 SELECT   *,
>                          ROW_NUMBER() OVER (PARTITION BY window_start, 
> window_end ORDER BY view_time DESC) AS rownum
>                 FROM     (
>                                   SELECT   window_start,
>                                            window_end,
>                                            product_id,
>                                            SUM(view_time) AS view_time,
>                                            COUNT(*)       AS cnt
>                                   FROM     TABLE(TUMBLE(TABLE 
> `shoe_clickstream`, DESCRIPTOR($rowtime), INTERVAL '10' MINUTES))
>                                   GROUP BY window_start,
>                                            window_end,
>                                            product_id))
> WHERE  rownum <= 5;{code}
>  
> With the feature disabled (default), the number of files in rocksdb working 
> directory (as well as in the checkpoint) should grow indefinitely.
>  
> With feature enabled, the number of files should stays constant (as they 
> should get merged with each other).
> To enable the feature, set 
> {code:java}
> state.backend.rocksdb.manual-compaction.min-interval{code}
>  set to 1 minute for example.
>  
> Please consult 
> [https://github.com/apache/flink/blob/e7d7db3b6f87e53d9bace2a16cf95e5f7a79087a/flink-state-backends/flink-statebackend-rocksdb/src/main/java/org/apache/flink/contrib/streaming/state/sstmerge/RocksDBManualCompactionOptions.java#L29]
>  for other options if necessary.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to