Mailing-List: contact cloudstack-users-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: cloudstack-users@incubator.apache.org
Received-SPF: pass (nike.apache.org: local policy)
Date: Thu, 25 Oct 2012 14:13:11 +0200 (CEST)
From: Julien Garet <julien.garet@inria.fr>
To: cloudstack-users@incubator.apache.org
Message-ID: <2100712438.29917693.1351167191248.JavaMail.root@inria.fr>
In-Reply-To: 
 <CAD3utPFQFGyfTrTL6W8yb2wUc2fQW=fEzyc6PXQ3=cmqMrxADA@mail.gmail.com>
Subject: Re: maximum hosts per cluster
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_Part_29917692_1069075320.1351167191247"

------=_Part_29917692_1069075320.1351167191247
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

> De: "JoonHyun Pac" <shibaboy@gmail.com>
> =C0: cloudstack-users@incubator.apache.org
> Envoy=E9: Jeudi 25 Octobre 2012 12:59:50
> Objet: Re: maximum hosts per cluster

> The limit of hosts are 16 but it's too big for just one master host.
> I'm using 8 hosts per cluster and it works fine.
What happens if you have 16 hosts ? What's the difference from the cloud co=
ntroller to have 16 hosts in one cluster or 16 hosts in two clusters of 8.=
=20

> On Thu, Oct 25, 2012 at 5:59 PM, Julien Garet <julien.garet@inria.fr>
> wrote:

> > Hello,
> >
> >
> > Still in deployment planning, I am wondering how much hosts should
> > I put
> > in a single cluster ? I am using Cloudstack 3.0.2 with KVM on
> > CentOS 6.2
> > hosts.
> >
> >
> > What are the advantages and drawbacks of putting hosts in the same
> > cluster
> > ?
> >
> >
> > From what I see, if I am in the same cluster, I can migrate
> > instances
> > between hosts, what I can not do between different clusters, I can
> > also
> > migrate primary storage.
> >
> >
> > What happens if I had more than 16 hosts in a KVM cluster (16 seems
> > to be
> > the limit from the advanced install guide).
> >
> >
> > I am using iSCSI luns as primary storage and Dell C6220 servers (4
> > servers
> > in 2U chassis, sharing power input). I thought first I would divide
> > each
> > chassis in its own cluster, but, from management side, I need to
> > create a
> > new lun for every chassis. If I need to do some operations on a
> > chassis, I
> > need to put all servers in the cluster in maintenance mode, but it
> > seems I
> > can't do that ! At least one server should be up, no ?
> >
> >
> > Any advice would be greatly appreciated
> >
> >
> > ------
> > Attention, changement d'adresse et de num=E9ro de t=E9l=E9phone !
> > Julien GARET
> > DSI / SESI
> > Bureau B201
> > Inria Lille-Nord Europe
> > B=E2timent B, 6 rue Heloise
> > 59650 Villeneuve d'Ascq
> > t=E9l=E9phone: 03.59.35.86.86
> > portable: 06.79.88.10.47
> >

------=_Part_29917692_1069075320.1351167191247--