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 Wed, 19 Oct 2016 18:53:18 GMT
Alex Behm has posted comments on this change.

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

Patch Set 3:

File be/src/exec/

Line 36: DEFINE_int32(kudu_mutation_buffer_size, 100 * 1024 * 1024, "The size (bytes) of the
> My understanding from the Kudu team is that we always want this enabled, an
That's fine, I was mostly concerned about those possible error cases in Open() where we'd
abort the query if we fail to set these buffer options in the session_ (and the user has no
way of disabling that behavior)

Shouldn't we check the value of this config option for sanity, e.g. 0 or negative values or
some really low number like 20.

Line 128:   KUDU_RETURN_IF_ERROR(session_->SetFlushMode(
> I don't think these configuration operations should fail unless we called t
Why not DCHECK if we are sure it's our bug?

Or is it possible that this call fails due to session timeout or similar?
File be/src/exec/kudu-table-sink.h:

Line 35: /// The data is added to a Kudu in Send(). The Kudu client is configured to automatically
typo: remove first 'a'

Line 37: /// requires specifying a mutation buffer size and a buffer flush watermark percentage.
... in the Kudu client.

Line 42: /// that results in Kudu flushing after 7MB is in a particular buffer (of the 100MB
is in a buffer for a particular destination?
File testdata/workloads/functional-query/queries/QueryTest/kudu_crud.test:

Line 323: ====
Don't we need to also modify the expected ERRORS of existing tests for KUDU_KEY_ALREADY_PRESENT
Or do we not have tests for that?

To view, visit
To unsubscribe, visit

Gerrit-MessageType: comment
Gerrit-Change-Id: I5542b9a061b01c543a139e8722560b1365f06595
Gerrit-PatchSet: 3
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