accumulo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Josh Elser <josh.el...@gmail.com>
Subject Re: Failed to connect to server
Date Fri, 28 Mar 2014 02:12:43 GMT
You may not be getting a login-shell (your shell's configuration file is 
not being loaded) when it SSHs to the nodes in the cluster which would 
cause the HADOOP_PREFIX problem you presented.

An easy way to circumvent this is to just ensure that the environment 
variables in accumulo-env.sh are properly set. This will be a guarantee 
that Accumulo will always have the correct environment variables set.

On 3/27/14, 6:57 PM, escribamarko wrote:
> Hi Sean,
>
> Thank you for the response.
>
> This is the host name:
>
> ec2-XXXXXXXX.us-west-1.compute.amazonaws.com
>
> Our instance is deployed in AWS. I've tried
>
> $> host ec2-XXXXXXXX.us-west-1.compute.amazonaws.com
>   ec2-XXXXXXXX.us-west-1.compute.amazonaws.com has address 172.30.1.78
>
> 172.30.1.78 is configured in master/slaves files, as well as on gc, monitor
> and tracer. This is the internal address on aws instance.
>
> When I've tried ec2-XXXXXXXX.us-west-1.compute.amazonaws.com on
> master/slaves files, and start the accumulo, I've got an error on logs:
>
> "You must set HADOOP_PREFIX"
>
> But echo $HADOOP_PREFIX is returning the correct installation directory of
> hadoop.
>
> On hadoop core-site.xml
>
> <name>fs.default.name</name>
> <value>hdfs://localhost:9000</value>
>
> Thanks in advance.
>
>
>
> --
> View this message in context: http://apache-accumulo.1065345.n5.nabble.com/Failed-to-connect-to-server-tp8484p8518.html
> Sent from the Developers mailing list archive at Nabble.com.
>

Mime
View raw message