Mailing-List: contact user-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@hbase.apache.org
Received-SPF: pass (athena.apache.org: domain of
 ramkrishna.s.vasudevan@gmail.com designates 209.85.216.45 as permitted
 sender)
MIME-Version: 1.0
In-Reply-To: 
 <CAP2fyVfyS9E+_uYYOfZ1aKBLP==nGoY_GgMcxC6bK7dPdSp8tw@mail.gmail.com>
References: 
 <CAP2fyVfyS9E+_uYYOfZ1aKBLP==nGoY_GgMcxC6bK7dPdSp8tw@mail.gmail.com>
Date: Tue, 11 Jun 2013 06:18:42 +0530
Message-ID: 
 <CAAT7Mkp3L1FmhTGqRG_H6oj7Q1Ab70CPM=R=TpZjNoUeu19bfA@mail.gmail.com>
Subject: Re: SSD vs Spinning disks
From: ramkrishna vasudevan <ramkrishna.s.vasudevan@gmail.com>
To: user@hbase.apache.org
Content-Type: multipart/alternative; boundary=001a11c305d03f3c5804ded6424f

--001a11c305d03f3c5804ded6424f
Content-Type: text/plain; charset=ISO-8859-1

Currently this feature is not available.  But there is a JIRA that has been
raised to help in achieving this.
https://issues.apache.org/jira/browse/HBASE-6572.  Feel free to give your
comments over there.

Regards
Ram


On Tue, Jun 11, 2013 at 6:14 AM, Lucas Stanley <lucas23145@gmail.com> wrote:

> Hi,
>
> If I understand HBase's architecture correctly, it is only the WAL that
> needs to be placed on a SSD to make writes perform better?
>
> Because a write comes to a region server, then goes to the WAL and
> memstore. If the WAL is on SSD on this node and 2 other DataNodes, then the
> append writes will finish much faster, right?
>
> So, is there a way in HDFS to tell it to only put the HBase WALs
> (/hbase/.logs/) on a SSD disk and to keep the other HFiles on traditional
> spinning disks?
>

--001a11c305d03f3c5804ded6424f--