accumulo-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sean Busbey <busbey+li...@cloudera.com>
Subject Re: Installing with Hadoop 2.2.0
Date Wed, 19 Mar 2014 01:19:22 GMT
On Tue, Mar 18, 2014 at 7:58 PM, Sean Busbey <busbey+lists@cloudera.com>wrote:

>
> On Mar 18, 2014 7:51 PM, "Benjamin Parrish" <benjamin.d.parrish@gmail.com>
> wrote:
> >
> > HADOOP_CONF_DIR=/usr/local/hadoop/etc/hadoop is set in all
> ~/.bash_profile files as needed.
> >
> >
>
> Can you add to the gist the output of running
>
> $> find $HADOOP_CONF_DIR
>
> As the user who runs the tablet server on the same host you ran the
> classpath command on?
>


Ah! the problem is that you're using .bash_profile to set up the env.
 bash_profile is only used for login shells.

Presuming you use the provided $ACCUMULO_HOME/bin/start-all.sh, it does not
use login shells when starting the remote processes.

You can verify this by doing

ssh ${HOST} "bash -c 'echo ${HADOOP_CONF_DIR:-no hadoop conf}'"

as the accumulo user on the master server, where HOST is the tserver.

You can fix this by either setting up the env for non-login (and pref
non-interactive) shells or by making sure your accumulo-env.sh properly
defaults to the right place.

-Sean

Mime
View raw message