Commit d6706af0 authored by Patrick D. Hunt's avatar Patrick D. Hunt
Browse files

ZOOKEEPER-145. write detailed release notes for users migrating from 2.x to 3.0

git-svn-id: https://svn.apache.org/repos/asf/hadoop/zookeeper/trunk@706839 13f79535-47bb-0310-9956-ffa450edef68
parent 75407982
......@@ -37,6 +37,9 @@ Backward compatibile changes:
BUGFIXES:
ZOOKEEPER-145. write detailed release notes for users migrating from 2.x
to 3.0 (phunt)
ZOOKEEPER-23. Auto reset of watches on reconnect (breed via phunt)
ZOOKEEPER-191. forrest docs for upgrade. (mahadev via phunt)
......
......@@ -204,6 +204,9 @@ document.write("Last Published: " + document.lastModified);
<li>
<a href="#migration_data">Migrating Server Data</a>
</li>
<li>
<a href="#migration_config">Migrating Server Configuration</a>
</li>
</ul>
</li>
<li>
......@@ -265,11 +268,24 @@ Note: ZooKeeper increments the major version number (major.minor.fix) when backw
<a href="#migration_data">Migrating Server Data</a>
</p>
</li>
<li>
<p>
<a href="#migration_config">Migrating Server Configuration</a>
</p>
</li>
</ul>
<a name="N1003F"></a><a name="migration_code"></a>
<a name="N10045"></a><a name="migration_code"></a>
<h3 class="h4">Migrating Client Code</h3>
<a name="N10045"></a><a name="Watch+Management"></a>
<p>
The underlying client-server protocol has changed in version 3.0.0
of ZooKeeper. As a result clients must be upgraded along with
serving clusters to ensure proper operation of the system (old
pre-3.0.0 clients are not guaranteed to operate against upgraded
3.0.0 servers and vice-versa).
</p>
<a name="N1004E"></a><a name="Watch+Management"></a>
<h4>Watch Management</h4>
<p>
In previous releases of ZooKeeper any watches registered by clients were lost if the client lost a connection to a ZooKeeper server.
......@@ -285,7 +301,7 @@ This release also allows clients to specify call specific watch functions.
This gives the developer the ability to modularize logic in different watch functions rather than cramming everything in the watch function attached to the ZooKeeper handle.
Call specific watch functions receive all session events for as long as they are active, but will only receive the watch callbacks for which they are registered.
</p>
<a name="N10052"></a><a name="Java+API"></a>
<a name="N1005B"></a><a name="Java+API"></a>
<h4>Java API</h4>
<ol>
......@@ -307,7 +323,7 @@ Call specific watch functions receive all session events for as long as they are
Also see <a href="http://hadoop.apache.org/zookeeper/docs/current/api/index.html">the current java API</a>
</p>
<a name="N10084"></a><a name="C+API"></a>
<a name="N1008D"></a><a name="C+API"></a>
<h4>C API</h4>
<ol>
......@@ -316,7 +332,7 @@ Also see <a href="http://hadoop.apache.org/zookeeper/docs/current/api/index.html
</li>
</ol>
<a name="N10097"></a><a name="migration_data"></a>
<a name="N100A0"></a><a name="migration_data"></a>
<h3 class="h4">Migrating Server Data</h3>
<p>
The following issues resulted in changes to the on-disk data format (the snapshot and transaction log files contained within the ZK data directory) and require a migration utility to be run.
......@@ -462,10 +478,31 @@ The following issues resulted in changes to the on-disk data format (the snapsho
</li>
</ol>
<a name="N10135"></a><a name="migration_config"></a>
<h3 class="h4">Migrating Server Configuration</h3>
<p>
There is a significant change to the ZooKeeper server configuration file.
</p>
<p>The default election algorithm, specified by
the <em>electionAlg</em> configuration attribute, has
changed from a default of <em>0</em> to a default
of <em>3</em>. See
<a href="zookeeperAdmin.html#sc_clusterOptions">Cluster
Options</a> section of the administrators guide, specifically
the <em>electionAlg</em>
and <em>server.X</em> properties.
</p>
<p>
You will either need to explicitly
set <em>electionAlg</em> to it's previous default value
of <em>0</em> or change
your <em>server.X</em> options to include the leader
election port.
</p>
</div>
<a name="N1012D"></a><a name="changes"></a>
<a name="N10162"></a><a name="changes"></a>
<h2 class="h3">Changes Since ZooKeeper 2.2.1</h2>
<div class="section">
<p>
......
This diff is collapsed.
......@@ -62,11 +62,20 @@ Note: ZooKeeper increments the major version number (major.minor.fix) when backw
<itemizedlist>
<listitem><para><ulink url="#migration_code">Migrating Client Code</ulink></para></listitem>
<listitem><para><ulink url="#migration_data">Migrating Server Data</ulink></para></listitem>
<listitem><para><ulink url="#migration_config">Migrating Server Configuration</ulink></para></listitem>
</itemizedlist>
<section id="migration_code">
<title>Migrating Client Code</title>
<para>
The underlying client-server protocol has changed in version 3.0.0
of ZooKeeper. As a result clients must be upgraded along with
serving clusters to ensure proper operation of the system (old
pre-3.0.0 clients are not guaranteed to operate against upgraded
3.0.0 servers and vice-versa).
</para>
<section>
<title>Watch Management</title>
......@@ -186,6 +195,34 @@ The following issues resulted in changes to the on-disk data format (the snapsho
</orderedlist>
</section>
<section id="migration_config">
<title>Migrating Server Configuration</title>
<para>
There is a significant change to the ZooKeeper server configuration file.
</para>
<para>The default election algorithm, specified by
the <emphasis>electionAlg</emphasis> configuration attribute, has
changed from a default of <emphasis>0</emphasis> to a default
of <emphasis>3</emphasis>. See
<ulink url="zookeeperAdmin.html#sc_clusterOptions">Cluster
Options</ulink> section of the administrators guide, specifically
the <emphasis>electionAlg</emphasis>
and <emphasis>server.X</emphasis> properties.
</para>
<para>
You will either need to explicitly
set <emphasis>electionAlg</emphasis> to it's previous default value
of <emphasis>0</emphasis> or change
your <emphasis>server.X</emphasis> options to include the leader
election port.
</para>
</section>
</section>
<section id="changes">
......
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment