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 colpclark@gmail.com designates
 209.85.223.182 as permitted sender)
Subject: Re: How safe is "nodetool move" in 1.2 ?
References: <lilr7m$2tb$1@ger.gmane.org>
 <CAOP8aV-6MRA6aJfE9dqbm3U0nE+MJzAyrS8+6C7DbpZ0XvBSaA@mail.gmail.com>
 <CAAwnuDtNvYNS_=ofiOQVauZ_2BZ_o72SgpijZmGfqCtFVmpMDg@mail.gmail.com>
From: Colin <colpclark@gmail.com>
Content-Type: multipart/alternative;
	boundary=Apple-Mail-5C69DDDA-9717-42E0-81FD-7AD536868292
In-Reply-To: 
 <CAAwnuDtNvYNS_=ofiOQVauZ_2BZ_o72SgpijZmGfqCtFVmpMDg@mail.gmail.com>
Message-Id: <2286E9E8-D403-4013-94F7-02F8308623B1@gmail.com>
Date: Wed, 16 Apr 2014 19:59:05 -0500
To: "user@cassandra.apache.org" <user@cassandra.apache.org>
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (1.0)


--Apple-Mail-5C69DDDA-9717-42E0-81FD-7AD536868292
Content-Type: text/plain;
	charset=us-ascii
Content-Transfer-Encoding: quoted-printable

I have recently tested this scenario under a couple versions of Cassandra an=
d have been able to write and read to/from the cluster while performing a mo=
ve.

I performed these tests utilizing an RF=3D2 on a three node cluster while pe=
rforming quorum reads and received no errors due to unavailable replicas.

I will be doing some more testing on this under different scenarios, but so f=
ar so good

However, I would strongly recommend an RF of at least 3 when performing quor=
um based reads because otherwise you're subject to failed reads in event of l=
osing one node.

--
Colin
320-221-9531


> On Apr 16, 2014, at 6:28 PM, Richard Low <richard@wentnet.com> wrote:
>=20
>> On 16 April 2014 05:08, Jonathan Lacefield <jlacefield@datastax.com> wrot=
e:
>> Assuming you have enough nodes not undergoing "move" to meet your CL requ=
irements, then yes, your cluster will still accept reads and writes.   Howev=
er, it's always good to test this before doing it in production to ensure yo=
ur cluster and app will function as designed.
>=20
> This is not a correctness requirement: writes go to the move source and de=
stination during the move and reads come from the source. Otherwise you coul=
d lose data during move (and certainly would lose data if replication factor=
 was one). However, nodes that are involved in the move will be slower so it=
 will be better for performance to not move nodes that share replicas simult=
aneously.
>=20
> Richard.

--Apple-Mail-5C69DDDA-9717-42E0-81FD-7AD536868292
Content-Type: text/html;
	charset=utf-8
Content-Transfer-Encoding: quoted-printable

<html><head><meta http-equiv=3D"content-type" content=3D"text/html; charset=3D=
utf-8"></head><body dir=3D"auto"><div>I have recently tested this scenario u=
nder a couple versions of Cassandra and have been able to write and read to/=
from the cluster while performing a move.</div><div><br></div><div>I perform=
ed these tests utilizing an RF=3D2 on a three node cluster while performing q=
uorum reads and received no errors due to unavailable replicas.</div><div><b=
r></div><div>I will be doing some more testing on this under different scena=
rios, but so far so good</div><div><br></div><div>However, I would strongly r=
ecommend an RF of at least 3 when performing quorum based reads because othe=
rwise you're subject to failed reads in event of losing one node.<br><br>--<=
div>Colin</div><div>320-221-9531</div><div><br></div></div><div><br>On Apr 1=
6, 2014, at 6:28 PM, Richard Low &lt;<a href=3D"mailto:richard@wentnet.com">=
richard@wentnet.com</a>&gt; wrote:<br><br></div><blockquote type=3D"cite"><d=
iv><div dir=3D"ltr"><div class=3D"gmail_extra"><div class=3D"gmail_quote">On=
 16 April 2014 05:08, Jonathan Lacefield <span dir=3D"ltr">&lt;<a href=3D"ma=
ilto:jlacefield@datastax.com" target=3D"_blank">jlacefield@datastax.com</a>&=
gt;</span> wrote:<br>

<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px=
 #ccc solid;padding-left:1ex"><div dir=3D"ltr">Assuming you have enough node=
s not undergoing "move" to meet your CL requirements, then yes, your cluster=
 will still accept reads and writes. &nbsp; However, it's always good to tes=
t this before doing it in production to ensure your cluster and app will fun=
ction as designed.</div>

</blockquote><div><br></div><div>This is not a correctness requirement: writ=
es go to the move source and destination during the move and reads come from=
 the source. Otherwise you could lose data during move (and certainly would l=
ose data if replication factor was one). However, nodes that are involved in=
 the move will be slower so it will be better for performance to not move no=
des that share replicas simultaneously.</div>

<div><br></div><div>Richard.</div></div></div></div>
</div></blockquote></body></html>=

--Apple-Mail-5C69DDDA-9717-42E0-81FD-7AD536868292--