Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@cassandra.apache.org
Received-SPF: pass (nike.apache.org: domain of shiyingjie1983@gmail.com
 designates 74.125.83.44 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:in-reply-to:references:date:message-id:subject:from:to
         :content-type;
        b=BqEwQkyC/OtisRQ7098h5FsyqTHpXz04EECXubtBw+2uCeZHpgvEUjlquw4GvDgPe1
         854x2vuLv/+zLy+roMjU1aDIvCtrtHQEYS/jxp7IzEsWkI3SfbKR5JYDxfYStAj3vAPL
         Z7BuB8Zogu2bvZIEiwUgA0JBlBY9FWte3Y6pM=
MIME-Version: 1.0
In-Reply-To: <AANLkTimBRqtiesVTh9gMFhRSZzcH4RpPNRsub_VeStZP@mail.gmail.com>
References: <AANLkTimabrUTnhOJrgfBR0m8IFJ1-Akly0au85VSeGVQ@mail.gmail.com>
	 <AANLkTimBRqtiesVTh9gMFhRSZzcH4RpPNRsub_VeStZP@mail.gmail.com>
Date: Fri, 21 May 2010 09:50:51 +0800
Message-ID: <AANLkTinskVt9mudY0JDSBe2EtPAMmCRnljb201PI2HSK@mail.gmail.com>
Subject: Re: What happened if one server involved in the process of data
	reading fail?
From: =?GB2312?B?yrfTor3c?= <shiyingjie1983@gmail.com>
To: user@cassandra.apache.org
Content-Type: multipart/alternative; boundary=0016e68ee1e8bd5693048710eb30

--0016e68ee1e8bd5693048710eb30
Content-Type: text/plain; charset=GB2312
Content-Transfer-Encoding: quoted-printable

What inner mechanism does Cassandra adopt to get this kind of fault
tolerance?

2010/5/20 Simon Smith <simongsmith@gmail.com>

> On Thu, May 20, 2010 at 8:08 AM, =CA=B7=D3=A2=BD=DC <shiyingjie1983@gmail=
.com> wrote:
> > Hi, All,
> >     I am now learning the mechanism Cassandra adopts to get high
> > availability and fault tolerance.  As I know, we should connect to one
> > server of Cassandra first, then we can read or write data  through it, =
so
> if
> > the server which we connect to get down, what will happen? Should we ha=
ve
> to
> > reconnect another server or will Cassandra control this situation?
>
>
> The approach we're taking is to put the software load-balancer haproxy
> in front of our cassandra cluster.  Use "mode tcp" within haproxy's
> config.  I notice that Tragedy (http://github.com/enki/tragedy/) also
> lets you put a list of servers into the connection call (we're going
> to put the list of haproxy load balancers here).
>
>
>
> > Another sutiation, if the server which is involved in the process of da=
ta
> reading
> > fail, what will Cassandra do?
>
>
> If you're using Thrift to connect, catch the exceptions that library
> throws if unable to connect and then try to connect again.   This is
> going to happen - if/when a node goes down it causes the entire
> cluster to hiccup a little, so if it is critical that any particular
> read transaction succeeds, you may need to sleep as much as 5 seconds
> (this is just my experience).
>
>
> >     Thanks a lot!
> >
> > Yingjie
>

--0016e68ee1e8bd5693048710eb30
Content-Type: text/html; charset=GB2312
Content-Transfer-Encoding: quoted-printable

What inner mechanism does Cassandra adopt to get this kind of fault toleran=
ce?<br><br>
<div class=3D"gmail_quote">2010/5/20 Simon Smith <span dir=3D"ltr">&lt;<a h=
ref=3D"mailto:simongsmith@gmail.com">simongsmith@gmail.com</a>&gt;</span><b=
r>
<blockquote class=3D"gmail_quote" style=3D"PADDING-LEFT: 1ex; MARGIN: 0px 0=
px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">
<div class=3D"im">On Thu, May 20, 2010 at 8:08 AM, =CA=B7=D3=A2=BD=DC &lt;<=
a href=3D"mailto:shiyingjie1983@gmail.com">shiyingjie1983@gmail.com</a>&gt;=
 wrote:<br>&gt; Hi, All,<br>&gt; &nbsp;&nbsp;&nbsp; I am now learning the m=
echanism Cassandra adopts to get high<br>
&gt; availability and fault tolerance.&nbsp; As I know, we should connect t=
o one<br>&gt; server of Cassandra first, then we can read or write data&nbs=
p; through it, so if<br>&gt; the server which we connect to get down, what =
will happen? Should we have to<br>
&gt; reconnect another server or will Cassandra control this situation?<br>=
<br><br></div>The approach we&#39;re taking is to put the software load-bal=
ancer haproxy<br>in front of our cassandra cluster. &nbsp;Use &quot;mode tc=
p&quot; within haproxy&#39;s<br>
config. &nbsp;I notice that Tragedy (<a href=3D"http://github.com/enki/trag=
edy/" target=3D"_blank">http://github.com/enki/tragedy/</a>) also<br>lets y=
ou put a list of servers into the connection call (we&#39;re going<br>to pu=
t the list of haproxy load balancers here).<br>

<div class=3D"im"><br><br><br>&gt; Another sutiation, if the server which i=
s involved in the process of data reading<br>&gt; fail,&nbsp;what will Cass=
andra do?<br><br><br></div>If you&#39;re using Thrift to connect, catch the=
 exceptions that library<br>
throws if unable to connect and then try to connect again. &nbsp; This is<b=
r>going to happen - if/when a node goes down it causes the entire<br>cluste=
r to hiccup a little, so if it is critical that any particular<br>read tran=
saction succeeds, you may need to sleep as much as 5 seconds<br>
(this is just my experience).<br><br><br>&gt; &nbsp;&nbsp;&nbsp;&nbsp;Thank=
s a lot!<br>&gt;<br>&gt; Yingjie<br></blockquote></div><br>

--0016e68ee1e8bd5693048710eb30--