Mailing-List: contact commits-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cassandra.apache.org
Date: Tue, 19 Sep 2017 10:22:00 +0000 (UTC)
From: "Aleksey Yeschenko (JIRA)" <jira@apache.org>
To: commits@cassandra.apache.org
Message-ID: <JIRA.13102665.1505503080000.145413.1505816520381@Atlassian.JIRA>
In-Reply-To: <JIRA.13102665.1505503080000@Atlassian.JIRA>
References: <JIRA.13102665.1505503080000@Atlassian.JIRA> <JIRA.13102665.1505503080493@jira-lw-us.apache.org>
Subject: [jira] [Commented] (CASSANDRA-13880) Fix short read protection for
 tables with no clustering columns
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
archived-at: Tue, 19 Sep 2017 10:22:07 -0000


    [ https://issues.apache.org/jira/browse/CASSANDRA-13880?page=3Dcom.atla=
ssian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=
=3D16171476#comment-16171476 ]=20

Aleksey Yeschenko commented on CASSANDRA-13880:
-----------------------------------------------

Thanks.

CircleCI run for [3.0|https://circleci.com/gh/iamaleksey/cassandra/34] has =
a bunch of annoying MV timeout issues again, and [dtests|https://builds.apa=
che.org/view/A-D/view/Cassandra/job/Cassandra-devbranch-dtest/326/testRepor=
t/] have those annoying git clone issues, again, but everything seems alrig=
ht otherwise.

Committed to 3.0 as [35e32f20ba8cba6cbfb1bee4252c0edd8684cdb1|https://githu=
b.com/apache/cassandra/commit/35e32f20ba8cba6cbfb1bee4252c0edd8684cdb1] and=
 merged with 3.11 and trunk. Dtest committed as [163f82c2db0e86d4dd8f312b29=
1ccd094891b986|https://github.com/apache/cassandra-dtest/commit/163f82c2db0=
e86d4dd8f312b291ccd094891b986].

> Fix short read protection for tables with no clustering columns
> ---------------------------------------------------------------
>
>                 Key: CASSANDRA-13880
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-1388=
0
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Coordination
>            Reporter: Aleksey Yeschenko
>            Assignee: Aleksey Yeschenko
>             Fix For: 3.0.15, 3.11.1
>
>
> CASSANDRA-12872 fixed counting replica rows, so that we do now fetch more=
 than one extra row if necessary.
> Fixing the issue caused consistency_test.py:TestConsistency.test_13747 to=
 start failing, by exposing a bug in the way we handle empty clusterings.
> When {{moreContents()}} asks for another row and {{lastClustering}} is {{=
EMPTY}}, the response again (and again) contains the row with {{EMPTY}} clu=
stering.
> SRP assumes it=E2=80=99s a new row, counts it as one, gets confused and k=
eeps asking for more, in a loop, again and again.
> Arguably, a response to a read command with the following non-inclusive {=
{ClusteringIndexFilter}}:
> {code}
> command.clusteringIndexFilter(partitionKey).forPaging(metadata.comparator=
, Clustering.EMPTY, false);
> {code}
> ... should return nothing at all rather than a row with an empty clusteri=
ng.
> Also arguably, SRP should not even attempt to fetch more rows if {{lastCl=
ustering =3D=3D Clustering.EMPTY}}. In a partition key only column
> we shouldn=E2=80=99t expect any more rows.
> This JIRA is to fix the latter issue on SRP side - to modify SRP logic to=
 short-circuit execution if {{lastClustering}} was an {{EMPTY}} one instead=
 of querying pointlessly for non-existent extra rows.


--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cassandra.apache.org
For additional commands, e-mail: commits-help@cassandra.apache.org