Mailing-List: contact dev-help@whirr.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@whirr.apache.org
Date: Wed, 14 Sep 2011 20:33:09 +0000 (UTC)
From: "Mingjie Lai (JIRA)" <jira@apache.org>
To: dev@whirr.apache.org
Message-ID: 
 <1721232105.28194.1316032389134.JavaMail.tomcat@hel.zones.apache.org>
In-Reply-To: 
 <1152549856.16147.1315807272535.JavaMail.tomcat@hel.zones.apache.org>
Subject: [jira] [Commented] (WHIRR-383) HBase cluster launch problem
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable


    [ https://issues.apache.org/jira/browse/WHIRR-383?page=3Dcom.atlassian.=
jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D13104=
859#comment-13104859 ]=20

Mingjie Lai commented on WHIRR-383:
-----------------------------------

Rok.

> I used hbase shell as a basic example I also tried connecting with ycsb f=
rom local machine and got a similar could not resolve error.

If you really want to run ycsb against hbase, you'd better start it within =
the same cluster of hbase. Since hbase client(ycsb here) will talk to each =
individual region server directly to get/put data. You don't want your loca=
l network to be the bottleneck for a performance test.=20

For hbase applications in general, it's also recommended to run client at t=
he same cluster to fully utilize the local network bandwidth to get best th=
roughput.=20

For us (we use ec2 + hbase, but not thru whirr), we usually start an applic=
ation server along with a hbase cluster, and this application server is use=
d for performance testing or any other hbase application.=20

Just FYI.

> HBase cluster launch problem
> ----------------------------
>
>                 Key: WHIRR-383
>                 URL: https://issues.apache.org/jira/browse/WHIRR-383
>             Project: Whirr
>          Issue Type: Bug
>          Components: service/hbase
>    Affects Versions: 0.6.0
>         Environment: Amazon EC2
>            Reporter: Rok Zlender
>
> I am trying to launch a hbase cluster with following settings
> {noformat}
> whirr.cluster-name=3Dhbase-0.90
> whirr.instance-templates=3D1 zookeeper+hadoop-namenode+hadoop-jobtracker+=
hbase-master,2 hadoop-datanode+hadoop-tasktracker+hbase-regionserver
> hbase-site.dfs.replication=3D1
> whirr.provider=3Daws-ec2
> whirr.identity=3D${env:AWS_ACCESS_KEY_ID}
> whirr.credential=3D${env:AWS_SECRET_ACCESS_KEY}
> whirr.hardware-id=3Dm1.large #Only change from HBase recipe that comes wi=
th 0.6.0
> whirr.image-id=3Dus-east-1/ami-da0cf8b3
> whirr.location-id=3Dus-east-1
> whirr.hbase.tarball.url=3Dhttp://apache.cu.be/hbase/hbase-0.90.3/hbase-0.=
90.3.tar.gz
> whirr.hadoop.tarball.url=3Dhttp://archive.cloudera.com/cdh/3/hadoop-0.20.=
2-cdh3u1.tar.gz
> {noformat}
> The only difference between my configuration and stock HBase recipe is th=
at I used m1.large ec2 instances. Cluster launches without problems and I c=
an connect to master node and do
> {noformat}
> rok@domU-12-31-39-15-24-20:/usr/local/hbase-0.90.3$ bin/hbase shell
> HBase Shell; enter 'help<RETURN>' for list of supported commands.
> Type "exit<RETURN>" to leave the HBase Shell
> Version 0.90.3, r1100350, Sat May  7 13:31:12 PDT 2011
> hbase(main):001:0> status
> 2 servers, 0 dead, 1.0000 average load
> hbase(main):002:0>=20
> {noformat}
> Web ui also shows that HBase 0.90.3, r1100350 with Hadoop 0.20.2-cdh3u1 w=
as started.
> But on my local machine from where I started this cluster I'm having prob=
lems connecting to it. I copied hbase-site.xml that whirr creates to hbase/=
conf folder and started hbase-proxy.sh. And when I try to connect and do st=
atus I get
> {noformat}
> =E2=9E=9C  hbase-0.90.3  bin/hbase shell
> HBase Shell; enter 'help<RETURN>' for list of supported commands.
> Type "exit<RETURN>" to leave the HBase Shell
> Version 0.90.3, r1100350, Sat May  7 13:31:12 PDT 2011
> hbase(main):001:0> status
> 11/09/11 10:49:43 ERROR hbase.HServerAddress: Could not resolve the DNS n=
ame of ip-10-2-38-254.ec2.internal:60000
> ERROR: java.lang.IllegalArgumentException: Could not resolve the DNS name=
 of ip-10-2-38-254.ec2.internal:60000
> Here is some help for this command:
> Show cluster status. Can be 'summary', 'simple', or 'detailed'. The
> default is 'summary'. Examples:
>   hbase> status
>   hbase> status 'simple'
>   hbase> status 'summary'
>   hbase> status 'detailed'
> hbase(main):002:0>=20
> {noformat}
> This looks similar to issue https://issues.apache.org/jira/browse/WHIRR-2=
01 but I don't know where to start debugging or tweaking to get it running =
properly.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira