impala-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alex Behm (Code Review)" <>
Subject [Impala-ASF-CR] IMPALA-4134,IMPALA-3704: Kudu INSERT improvements
Date Tue, 18 Oct 2016 22:07:15 GMT
Alex Behm has posted comments on this change.

Change subject: IMPALA-4134,IMPALA-3704: Kudu INSERT improvements

Patch Set 2:

File be/src/exec/

Line 36: DEFINE_int32(kudu_mutation_buffer_size, 100 * 1024 * 1024, "The size (bytes) of the
Can the user disable this or do we always want to have it enabled?

Line 128:   KUDU_RETURN_IF_ERROR(session_->SetFlushMode(
When would we hit these error conditions? Does the user have a recourse or will the insert/update/delete
just never work?

Line 133:   // The user specifies the total amount of memory used for buffering, but we don't
Please add a comment to the class header describing how the buffering works. I don't really
understand how these buffers and parameters play together.

For example, why would we just not use the flush watermark and decrease the mutation buffer
size instead.

Line 275:   // then PK conflicts are not treated as errors. Only the first real error is returned
Why would not use our conventional RuntimeState::LogError() mechanism with a proper error
code? Those will be deduped automatically and have the advantage of telling the user how many
duplicates were actually encountered.

To view, visit
To unsubscribe, visit

Gerrit-MessageType: comment
Gerrit-Change-Id: I5542b9a061b01c543a139e8722560b1365f06595
Gerrit-PatchSet: 2
Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-Owner: Matthew Jacobs <>
Gerrit-Reviewer: Alex Behm <>
Gerrit-Reviewer: Matthew Jacobs <>
Gerrit-Reviewer: Todd Lipcon <>
Gerrit-HasComments: Yes

View raw message