Mailing-List: contact solr-user-help@lucene.apache.org; run by ezmlm
Precedence: bulk
Reply-To: solr-user@lucene.apache.org
From: Toke Eskildsen <te@statsbiblioteket.dk>
To: solr_user lucene_apache <solr-user@lucene.apache.org>
Subject: Re: Whether SolrCloud can support 2 TB data?
Thread-Topic: Whether SolrCloud can support 2 TB data?
Thread-Index: AQHSFd1SXziuzYmuxE+aZ5u6US69pKCHrE0v///zuACAAHIBAIAAfOOAgABUEGs=
Date: Sat, 24 Sep 2016 17:48:50 +0000
Message-ID: <1474739330118.39784@statsbiblioteket.dk>
References: <1474662805089-4297790.post@n3.nabble.com>
	<CA+8QczC14irCJHbHhH=YCE+oSD0-8oHhvzTq5Nu1nub18tq7EQ@mail.gmail.com>
	<1474664334795-4297800.post@n3.nabble.com>
	<CAEa-FXgx6AkOJXvgg3fOM6OWz=7wK6+aUmMaXKVPLwYJd-whkw@mail.gmail.com>
	<CAEa-FXjriwHU9-yi83NG8TYcMthC8iybVo4_=DPP98EtZO6LQw@mail.gmail.com>
	<CAEa-FXgL0MS2n30z_xmZqfzHrf7px-=U=13NoH-osZaSQoJ2mQ@mail.gmail.com>
	<27d64747-4417-4c35-8c4e-d8028060a4fa@Spark>
	<CAGX0URjjZOMjPk=ToA--zVMAyLxBnp2f6C-bJrq6TWuHuYtfrA@mail.gmail.com>,<e74o1onbpvs334ppl0xtsyml0-2147483647@mailer.nylas.com>
In-Reply-To: <e74o1onbpvs334ppl0xtsyml0-2147483647@mailer.nylas.com>
Accept-Language: en-GB, da-DK, en-US
Content-Language: en-GB
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
archived-at: Sat, 24 Sep 2016 17:48:56 -0000

Regarding a 12TB index:=0A=
=0A=
Yago Riveiro <yago.riveiro@gmail.com> wrote:=0A=
=0A=
> Our cluster is small for the data we hold (12 machines with SSD and 32G o=
f=0A=
> RAM), but we don't need sub-second queries, we need facet with high=0A=
> cardinality (in worst case scenarios we aggregate 5M unique string values=
)=0A=
=0A=
> In a peak of inserts we can handle around 25K docs per second without any=
=0A=
> issue with 2 replicas and without compromise reads or put a node in stres=
s.=0A=
> Nodes in stress can eject him selfs from the Zookepeer cluster due a GC o=
r a=0A=
> lack of CPU to communicate.=0A=
=0A=
I am surprised that you manage to have this working on that hardware. As yo=
u have replicas, it seems to me that you handle 2*12TB of index with 12*32G=
B of RAM? This is very close to our setup (22TB of index with 320GB of RAM =
(updated last week from 256GB) per machine), but we benefit hugely from hav=
ing a static index.=0A=
=0A=
I assume the SSDs are local? How much memory do you use for heap on each ma=
chine?=0A=
=0A=
- Toke Eskildsen=0A=