Edit of 'Getting Started' to try and improve zookeeper section

git-svn-id: https://svn.apache.org/repos/asf/hadoop/hbase/trunk@786376 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Michael Stack 2009-06-19 05:38:51 +00:00
parent 26559442b7
commit cfc84a624f
2 changed files with 18 additions and 13 deletions

View File

@ -868,7 +868,7 @@ public class HTable {
* at a specified timestamp * at a specified timestamp
* *
* @param row row key * @param row row key
* @param timestamp timestamp * @param ts timestamp
* @param numVersions number of versions to return * @param numVersions number of versions to return
* @return RowResult is <code>null</code> if row does not exist. * @return RowResult is <code>null</code> if row does not exist.
* @throws IOException * @throws IOException
@ -2071,4 +2071,4 @@ public class HTable {
}; };
} }
} }
} }

View File

@ -178,18 +178,9 @@ The ZooKeeper configuration file for HBase is stored at <code>${HBASE_HOME}/conf
See the ZooKeeper <a href="http://hadoop.apache.org/zookeeper/docs/current/zookeeperStarted.html"> Getting Started Guide</a> See the ZooKeeper <a href="http://hadoop.apache.org/zookeeper/docs/current/zookeeperStarted.html"> Getting Started Guide</a>
for information about the format and options of that file. Specifically, look at the for information about the format and options of that file. Specifically, look at the
<a href="http://hadoop.apache.org/zookeeper/docs/current/zookeeperStarted.html#sc_RunningReplicatedZooKeeper">Running Replicated ZooKeeper</a> section. <a href="http://hadoop.apache.org/zookeeper/docs/current/zookeeperStarted.html#sc_RunningReplicatedZooKeeper">Running Replicated ZooKeeper</a> section.
After configuring <code>zoo.cfg</code>, in <code>${HBASE_HOME}/conf/hbase-env.sh</code>,
set the following to tell HBase to STOP managing its instance of ZooKeeper.
<blockquote>
<pre>
...
# Tell HBase whether it should manage it's own instance of Zookeeper or not.
export HBASE_MANAGES_ZK=false
</pre>
</blockquote>
</p> </p>
<p> <p>
Though not recommended, it can be convenient having HBase continue to manage Though not recommended, it can be convenient having HBase continue to manage
ZooKeeper even when in distributed mode (It can be good when testing or taking ZooKeeper even when in distributed mode (It can be good when testing or taking
@ -216,6 +207,20 @@ Be aware that this option is only recommanded for testing purposes as a failure
on that node would render HBase <b>unusable</b>. on that node would render HBase <b>unusable</b>.
</p> </p>
<p>
To tell HBase to stop managing a ZooKeeper instance, after configuring
<code>zoo.cfg</code> to point at the ZooKeeper Quorum you'd like HBase to
use, in <code>${HBASE_HOME}/conf/hbase-env.sh</code>,
set the following to tell HBase to STOP managing its instance of ZooKeeper.
<blockquote>
<pre>
...
# Tell HBase whether it should manage it's own instance of Zookeeper or not.
export HBASE_MANAGES_ZK=false
</pre>
</blockquote>
</p>
<p>Of note, if you have made <i>HDFS client configuration</i> on your hadoop cluster, HBase will not <p>Of note, if you have made <i>HDFS client configuration</i> on your hadoop cluster, HBase will not
see this configuration unless you do one of the following: see this configuration unless you do one of the following:
<ul> <ul>