Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@cassandra.apache.org
Received-SPF: pass (athena.apache.org: domain of strebkov@gmail.com designates
 209.85.192.46 as permitted sender)
Date: Wed, 04 Mar 2015 10:06:57 -0800 (PST)
MIME-Version: 1.0
Message-Id: <1425492417697.a19cfa28@Nodemailer>
In-Reply-To: 
 <CAH3f1B8F8anQTG1-U_eAYUkZmsROArL7SLVsnYRM4DDGer+gew@mail.gmail.com>
References: 
 <CAH3f1B8F8anQTG1-U_eAYUkZmsROArL7SLVsnYRM4DDGer+gew@mail.gmail.com>
From: "Mikhail Strebkov" <strebkov@gmail.com>
To: user@cassandra.apache.org
Cc: user@cassandra.apache.org
Subject: Re: Inconsistent count(*) and distinct results from Cassandra
Content-Type: multipart/alternative;
 boundary="----Nodemailer-0.5.0-?=_1-1425492417871"

------Nodemailer-0.5.0-?=_1-1425492417871
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

We have observed the same issue in our production Cassandra cluster (5 =
nodes in one DC). We use Cassandra 2.1.3 (I joined the list too late to =
realize we shouldn=E2=80=99t user 2.1.x yet) on Amazon machines (created =
from community AMI).


In addition to count variations with 5 to 10% we observe variations for the=
 query =E2=80=9Cselect * from table1 where time > '$fromDate' and time < =
'$toDate' allow filtering=E2=80=9D results. We iterated through the results=
 multiple times using official Java driver. We used that query for a huge =
data migration and were unpleasantly surprised that it is unreliable. In =
our case =E2=80=9Cnodetool repair=E2=80=9D didn=E2=80=99t fix the issue.


So I echo Frens questions.


Thanks,

Mikhail

On Wed, Mar 4, 2015 at 3:55 AM, Rumph, Frens Jan <mail@frensjan.nl> wrote:

> Hi,
> Is it to be expected that select count(*) from ... and select distinct
> partition-key-columns from ... to yield inconsistent results between
> executions even though the table at hand isn't written to=3F
> I have a table in a keyspace with replication=5Ffactor =3D 1 which is =
something
> like:
> CREATE TABLE tbl (
>     id frozen<id=5Ftype>,
>     bucket bigint,
>     offset int,
>     value double,
>     PRIMARY KEY ((id, bucket), offset)
> )
> The frozen udt is:
> CREATE TYPE id=5Ftype (
>     tags map<text, text>
> );
> When I do select count(*) from tbl several times the actual count varies
> with 5 to 10%. Also when performing select distinct id, bucket from tbl =
the
> results aren't consistent over several query executions. The table is =
not
> being written to at the time I performed the queries.
> Is this to be expected=3F Or is this a bug=3F Is there a alternative =
method /
> workaround=3F
> I'm using cqlsh 5.0.1 with Cassandra 2.1.2 on 64bit fedora 21 with =
Oracle
> Java 1.8.0=5F31.
> Thanks in advance,
> Frens Jan
------Nodemailer-0.5.0-?=_1-1425492417871
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable


<span id=3D=22mailbox-conversation=22><div>We have observed the same issue =
in our production Cassandra cluster (5 nodes in one DC). We use Cassandra 2=
.1.3 (I joined the list too late to realize we shouldn=E2=80=99t user 2.1.x=
 yet) on Amazon machines (created from community AMI).</div>
<div><br></div>
<div>In addition to count variations with 5 to 10% we observe variations =
for the query =E2=80=9Cselect * from table1 where time &gt; '$fromDate' and=
 time &lt; '$toDate' allow filtering=E2=80=9D results. We iterated through =
the results multiple times using official Java driver. We used that query =
for a huge data migration and were unpleasantly surprised that it is =
unreliable. In our case =E2=80=9Cnodetool repair=E2=80=9D didn=E2=80=99t =
fix the issue.</div>
<div><br></div>
<div>So I echo Frens questions.</div>
<div><br></div>
<div>Thanks,</div>
<div>Mikhail</div></span><div class=3D=22mailbox=5Fsignature=22>
<br><br></div>
<br><br><div class=3D=22gmail=5Fquote=22><p>On Wed, Mar 4, 2015 at 3:55 AM,=
 Rumph, Frens Jan <span dir=3D=22ltr=22>&lt;<a href=3D=22mailto:mail@frensj=
an.nl=22 target=3D=22=5Fblank=22>mail@frensjan.nl</a>&gt;</span> =
wrote:<br></p><blockquote class=3D=22gmail=5Fquote=22 style=3D=22margin:0 0=
 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;=22><div><div =
dir=3D=22ltr=22>Hi,<div><br></div>
<div>Is it to be expected that select count(*) from ... and select distinct=
 partition-key-columns from ... to yield inconsistent results between =
executions even though the table at hand isn't written to=3F</div>
<div><br></div>
<div>I have a table in a keyspace with replication=5Ffactor =3D 1 which is =
something like:</div>
<div><br></div>
<div>
<div>CREATE TABLE tbl (</div>
<div>=C2=A0 =C2=A0 id frozen&lt;id=5Ftype&gt;,</div>
<div>=C2=A0 =C2=A0 bucket bigint,</div>
<div>=C2=A0 =C2=A0 offset int,</div>
<div>=C2=A0 =C2=A0 value double,</div>
<div>=C2=A0 =C2=A0 PRIMARY KEY ((id, bucket), offset)</div>
<div>)</div>
</div>
<div><br></div>
<div>The frozen udt is:</div>
<div><br></div>
<div>
<div>CREATE TYPE id=5Ftype (</div>
<div>=C2=A0 =C2=A0 tags map&lt;text, text&gt;</div>
<div>);</div>
</div>
<div><br></div>
<div>When I do select count(*) from tbl several times the actual count =
varies with 5 to 10%. Also when performing select distinct id, bucket from =
tbl the results aren't consistent over several query executions. The table =
is not being written to at the time I performed the queries.</div>
<div><br></div>
<div>Is this to be expected=3F Or is this a bug=3F Is there a alternative =
method / workaround=3F</div>
<div><br></div>
<div>I'm using cqlsh 5.0.1 with Cassandra 2.1.2 on 64bit fedora 21 with =
Oracle Java 1.8.0=5F31.</div>
<div><br></div>
<div>Thanks in advance,</div>
<div>Frens Jan</div>
</div></div></blockquote></div><br>
------Nodemailer-0.5.0-?=_1-1425492417871--