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 jeremiah.jordan@gmail.com
 designates 209.85.214.178 as permitted sender)
From: Jeremiah D Jordan <jeremiah.jordan@gmail.com>
Content-Type: multipart/alternative;
 boundary="Apple-Mail=_EA978BF7-E0D3-406D-9FF9-690D40082187"
Message-Id: <89CC2836-0738-4F31-9D91-5B93EFA3BD5D@gmail.com>
Mime-Version: 1.0 (Mac OS X Mail 6.6 \(1510\))
Subject: Re: Virtual node support for Hadoop workloads
Date: Fri, 18 Oct 2013 10:36:15 -0500
References: 
 <CAKaZCX7pp6cTX488yYptQU8zzxvVt_cXjbZwcbnyNDP4upCzKQ@mail.gmail.com>
To: user@cassandra.apache.org
In-Reply-To: 
 <CAKaZCX7pp6cTX488yYptQU8zzxvVt_cXjbZwcbnyNDP4upCzKQ@mail.gmail.com>


--Apple-Mail=_EA978BF7-E0D3-406D-9FF9-690D40082187
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=iso-8859-1

Paulo,
If you have large data sizes then the vnodes with hadoop issue is moot.  =
You will get that many splits with or without vnodes.  The issues come =
when you don't have a lot of data, so all the extra splits slow =
everything down to a crawl because there are 256 times as many tasks =
created as you actually needed for your job.

So for large data sets, there is no issue.  For small data sets, you can =
run jobs, they will just be slower than if you didn't have vnodes.

-Jeremiah

On Oct 17, 2013, at 3:49 PM, Paulo Motta <pauloricardomg@gmail.com> =
wrote:

> Hello,
>=20
> According to DSE3.1 documentation [1], "DataStax recommends using =
virtual nodes only on data centers running purely Cassandra workloads. =
You should disable virtual nodes on data centers running either Hadoop =
or Solr workloads by setting num_tokens to 1.".
>=20
> There was a thread in this mailing list earlier this year [2], where =
it was suggested a workaround to the problem of having a minimum of one =
map task per token (unfeasible with vnodes). This suggestion involved =
implementing a new Hadoop InputSplitFormat that could combine many =
tokens from a single node, thus reducing the overhead of having too many =
tasks per node.=20
>=20
> Is there any JIRA ticket around this issue yet, or something being =
worked on to support VNodes for Hadoop workloads, or the suggestion =
remains to avoid VNodes for analytics workloads (hadoop, solr)?
>=20
> Thanks,=20
>=20
> --=20
> Paulo
>=20
> [1] =
http://www.datastax.com/docs/datastax_enterprise3.1/deploy/configuring_rep=
lication
> [2] =
http://mail-archives.apache.org/mod_mbox/cassandra-user/201302.mbox/%3CCAJ=
V_UYdqYmfStn5OetWrozQqbi+-yP3X-Ew9xtW=3DQY=3D2zGYDMA@mail.gmtokenail.com%3=
E


--Apple-Mail=_EA978BF7-E0D3-406D-9FF9-690D40082187
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html;
	charset=iso-8859-1

<html><head><meta http-equiv=3D"Content-Type" content=3D"text/html =
charset=3Diso-8859-1"></head><body style=3D"word-wrap: break-word; =
-webkit-nbsp-mode: space; -webkit-line-break: after-white-space; =
">Paulo,<div>If you have large data sizes then the vnodes with hadoop =
issue is moot. &nbsp;You will get that many splits with or without =
vnodes. &nbsp;The issues come when you don't have a lot of data, so all =
the extra splits slow everything down to a crawl because there are 256 =
times as many tasks created as you actually needed for your =
job.</div><div><br></div><div>So for large data sets, there is no issue. =
&nbsp;For small data sets, you can run jobs, they will just be slower =
than if you didn't have =
vnodes.</div><div><br></div><div>-Jeremiah</div><div><br><div><div>On =
Oct 17, 2013, at 3:49 PM, Paulo Motta &lt;<a =
href=3D"mailto:pauloricardomg@gmail.com">pauloricardomg@gmail.com</a>&gt; =
wrote:</div><br class=3D"Apple-interchange-newline"><blockquote =
type=3D"cite"><div dir=3D"ltr">Hello,<div><br></div><div>According to =
DSE3.1 documentation [1], "DataStax recommends using virtual nodes only =
on data centers running purely Cassandra workloads. You should disable =
virtual nodes on data centers running either Hadoop or Solr workloads by =
setting num_tokens to 1.".</div>

<div><br></div><div>There was a thread in this mailing list earlier this =
year [2], where it was suggested a workaround to the problem of having a =
minimum of one map task per token (unfeasible with vnodes). This =
suggestion involved implementing a new Hadoop InputSplitFormat that =
could combine many tokens from a single node, thus reducing the overhead =
of having too many tasks per node.&nbsp;</div>

<div><br></div><div>Is there any JIRA ticket around this issue yet, or =
something being worked on to support VNodes for Hadoop workloads, or the =
suggestion remains to avoid VNodes for analytics workloads (hadoop, =
solr)?</div>

<div><br></div><div>Thanks,&nbsp;<br clear=3D"all"><div><br></div>-- =
<br><div>Paulo</div><div><br></div>[1]&nbsp;<a =
href=3D"http://www.datastax.com/docs/datastax_enterprise3.1/deploy/configu=
ring_replication">http://www.datastax.com/docs/datastax_enterprise3.1/depl=
oy/configuring_replication</a><div>

<span style=3D"font-family:arial,sans-serif;line-height:15px"><i =
style=3D"font-style:normal"></i></span></div>
</div><div>[2]&nbsp;<a =
href=3D"http://mail-archives.apache.org/mod_mbox/cassandra-user/201302.mbo=
x/%3CCAJV_UYdqYmfStn5OetWrozQqbi+-yP3X-Ew9xtW=3DQY=3D2zGYDMA@mail.gmail.co=
m%3E">http://mail-archives.apache.org/mod_mbox/cassandra-user/201302.mbox/=
%3CCAJV_UYdqYmfStn5OetWrozQqbi+-yP3X-Ew9xtW=3DQY=3D2zGYDMA@mail.gmtokenail=
.com%3E</a></div>

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

--Apple-Mail=_EA978BF7-E0D3-406D-9FF9-690D40082187--