You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@impala.apache.org by "Alexey Serbin (Code Review)" <ge...@cloudera.org> on 2026/10/01 03:59:37 UTC

[Impala-ASF-CR] IMPALA-15463: Use Kudu live rows

Alexey Serbin has posted comments on this change. ( http://gerrit.cloudera.org:8080/24976 )

Change subject: IMPALA-15463: Use Kudu live rows
......................................................................


Patch Set 4: Code-Review+1

(4 comments)

http://gerrit.cloudera.org:8080/#/c/24976/4//COMMIT_MSG
Commit Message:

http://gerrit.cloudera.org:8080/#/c/24976/4//COMMIT_MSG@7
PS4, Line 7: rows
nit: row count stats?


http://gerrit.cloudera.org:8080/#/c/24976/4//COMMIT_MSG@12
PS4, Line 12: Adds a RESULTS:RETRY=N section to allow retrying queries while waiting
            : on asynchronous Kudu operations.
nit: perhaps, this might be a separate changelist on its own, but if you think it's better to come along with Kudu table stats, it's totally fine with me


http://gerrit.cloudera.org:8080/#/c/24976/4/fe/src/main/java/org/apache/impala/planner/KuduScanNode.java
File fe/src/main/java/org/apache/impala/planner/KuduScanNode.java:

PS4: 
BTW, recently per-tablet stats have been added in Kudu: http://gerrit.cloudera.org:8080/24650

Could Impala benefit from using that info?  I don't think it's exposed in the Java API yet, but it shouldn't be a big deal to add the required functionality, if needed.


http://gerrit.cloudera.org:8080/#/c/24976/4/testdata/workloads/functional-query/queries/QueryTest/kudu_stats.test
File testdata/workloads/functional-query/queries/QueryTest/kudu_stats.test:

http://gerrit.cloudera.org:8080/#/c/24976/4/testdata/workloads/functional-query/queries/QueryTest/kudu_stats.test@24
PS4, Line 24: # Kudu takes time to commit the rows and update its live row count.
            : select sleep(1000);
nit: I guess the required duration of the time interval depends on the TS-->master heartbeat period defined by the --heartbeat_interval_ms flag (default is 1000); setting the interval exactly at 1000 ms might lead to flakiness in rare cases, and I'd consider adding a bit of margin here.  Also, if Impala uses custom --heartbeat_interval_ms setting for Kudu in the test context, maybe use it instead of the hard-coded default value.



-- 
To view, visit http://gerrit.cloudera.org:8080/24976
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I17d806fa9f849bb66fac5dfb500f83fc6d115b58
Gerrit-Change-Number: 24976
Gerrit-PatchSet: 4
Gerrit-Owner: Michael Smith <mi...@cloudera.com>
Gerrit-Reviewer: Alexey Serbin <al...@apache.org>
Gerrit-Reviewer: David Rorke <dr...@cloudera.com>
Gerrit-Reviewer: Impala Public Jenkins <im...@cloudera.com>
Gerrit-Reviewer: Michael Smith <mi...@cloudera.com>
Gerrit-Reviewer: Wenzhe Zhou <wz...@cloudera.com>
Gerrit-Reviewer: Zoltan Borok-Nagy <bo...@cloudera.com>
Gerrit-Comment-Date: Thu, 01 Oct 2026 03:59:37 +0000
Gerrit-HasComments: Yes