Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@cassandra.apache.org
MIME-Version: 1.0
In-Reply-To: 
 <CABNXB2Ctx10=SZLpoNpLKjWtNsZHhAb-PQ_na01p6WjUJJPFGA@mail.gmail.com>
References: <5559BFA7.1080108@googlemail.com>
 <CABNXB2Ak+_4swnSTtXNW_1iSLRcZERnDpK6zTtvXg30twwwBdA@mail.gmail.com>
 <555AD0DF.4040106@googlemail.com>
 <CABNXB2Ctx10=SZLpoNpLKjWtNsZHhAb-PQ_na01p6WjUJJPFGA@mail.gmail.com>
From: Sylvain Lebresne <sylvain@datastax.com>
Date: Tue, 19 May 2015 10:04:51 +0200
Message-ID: 
 <CAKkz8Q3Xb7iVDienBy_=Ur68=J0x7CS1hL+b7VRWvpWezwNDzA@mail.gmail.com>
Subject: Re: Batch isolation within a single partition
To: "user@cassandra.apache.org" <user@cassandra.apache.org>
Content-Type: multipart/alternative; boundary=001a1133c1ba10547405166ac541

--001a1133c1ba10547405166ac541
Content-Type: text/plain; charset=UTF-8

On Tue, May 19, 2015 at 9:42 AM, DuyHai Doan <doanduyhai@gmail.com> wrote:

> If RF > 1, the consistency level at QUORUM cannot guarantee strict
> isolation (for normal mutation or batch). If you look at this slide:
> http://www.slideshare.net/doanduyhai/cassandra-introduction-apache-con-2014-budapest/25,
> you can see that the mutation is sent by the coordinator, in parallel, to
> all replicas.
>
>  Now it is very possible that due to network latency, the mutation is
> applied on the first replica and is applied with "some delay" (which can be
> at the order of microseconds) on other replicas.
>
>  Theoretically, one client can read updated value on first replica and old
> value on the other replicas, even at QUORUM.
>

Unfortunately different people will tend to have different definitions of
isolation and I don't seem to have the same definition than you but still,
I don't understand what you're talking about. Of course replicas might not
get a mutation at the same time, and yes, a read at QUORUM may thus not see
the most up to date value from all replicas. But the coordinator resolves
all responses together and return only the most recent one, so that doesn't
matter to the client and I don't see how that has anything to do with
isolation from the client perspective.

My response to the original question is that if by isolation you mean "can
a reader observe a write only partially applied", then for single partition
writes, Cassandra do offer isolation. One caveat however is that if 2
writes conflits, they are resolved using their timestamp and if the
timestamp are the same, resolution is based on values, which is not
necessarily intuitive and may make it sound like the writes where not
applied in isolation (even though technically they are), see
https://issues.apache.org/jira/browse/CASSANDRA-6123 for details on that
later problem. I'll note that my definition of isolation does not mean you
can't read stale data, and you can indeed if you use weak consistency
levels.

If you mean something else by isolation, then I think agreeing first on the
definition would be wise.

--
Sylvain

--001a1133c1ba10547405166ac541
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">On Tue, May 19, 2015 at 9:42 AM, DuyHai Doan <span dir=3D"=
ltr">&lt;<a href=3D"mailto:doanduyhai@gmail.com" target=3D"_blank">doanduyh=
ai@gmail.com</a>&gt;</span> wrote:<br><div class=3D"gmail_extra"><div class=
=3D"gmail_quote"><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px =
0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir=
=3D"ltr">If RF &gt; 1, the consistency level at QUORUM cannot guarantee str=
ict isolation (for normal mutation or batch). If you look at this slide:=C2=
=A0<a href=3D"http://www.slideshare.net/doanduyhai/cassandra-introduction-a=
pache-con-2014-budapest/25" target=3D"_blank">http://www.slideshare.net/doa=
nduyhai/cassandra-introduction-apache-con-2014-budapest/25</a>, you can see=
 that the mutation is sent by the coordinator, in parallel, to all replicas=
.<div><br></div><div>=C2=A0Now it is very possible that due to network late=
ncy, the mutation is applied on the first replica and is applied with &quot=
;some delay&quot; (which can be at the order of microseconds) on other repl=
icas.</div><div><br></div><div>=C2=A0Theoretically, one client can read upd=
ated value on first replica and old value on the other replicas, even at QU=
ORUM.</div></div></blockquote><div><br></div><div>Unfortunately different p=
eople will tend to have different definitions of isolation and I don&#39;t =
seem to have the same definition than you but still, I don&#39;t understand=
 what you&#39;re talking about. Of course replicas might not get a mutation=
 at the same time, and yes, a read at QUORUM may thus not see the most up t=
o date value from all replicas. But the coordinator resolves all responses =
together and return only the most recent one, so that doesn&#39;t matter to=
 the client and I don&#39;t see how that has anything to do with isolation =
from the client perspective.<br><br></div><div>My response to the original =
question is that if by isolation you mean &quot;can a reader observe a writ=
e only partially applied&quot;, then for single partition writes, Cassandra=
 do offer isolation. One caveat however is that if 2 writes conflits, they =
are resolved using their timestamp and if the timestamp are the same, resol=
ution is based on values, which is not necessarily intuitive and may make i=
t sound like the writes where not applied in isolation (even though technic=
ally they are), see <a href=3D"https://issues.apache.org/jira/browse/CASSAN=
DRA-6123">https://issues.apache.org/jira/browse/CASSANDRA-6123</a> for deta=
ils on that later problem. I&#39;ll note that my definition of isolation do=
es not mean you can&#39;t read stale data, and you can indeed if you use we=
ak consistency levels.<br><br></div><div>If you mean something else by isol=
ation, then I think agreeing first on the definition would be wise.<br><br>=
--<br></div><div>Sylvain<br></div></div></div></div>

--001a1133c1ba10547405166ac541--