HDFS-6806. HDFS Rolling upgrade document should mention the versions available. Contributed by J.Andreina.

(cherry picked from commit 82db3341bf)

Conflicts:
	hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt
This commit is contained in:
Akira Ajisaka 2015-03-10 11:22:11 +09:00
parent 6527be374b
commit c9a2e6f5db
2 changed files with 6 additions and 1 deletions

View File

@ -417,6 +417,9 @@ Release 2.7.0 - UNRELEASED
HDFS-7411. Change decommission logic to throttle by blocks rather than HDFS-7411. Change decommission logic to throttle by blocks rather than
nodes in each interval. (Andrew Wang via cdouglas) nodes in each interval. (Andrew Wang via cdouglas)
HDFS-6806. HDFS Rolling upgrade document should mention the versions
available. (J.Andreina via aajisaka)
OPTIMIZATIONS OPTIMIZATIONS
HDFS-7454. Reduce memory footprint for AclEntries in NameNode. HDFS-7454. Reduce memory footprint for AclEntries in NameNode.

View File

@ -47,7 +47,9 @@
These two capabilities make it feasible to upgrade HDFS without incurring HDFS downtime. These two capabilities make it feasible to upgrade HDFS without incurring HDFS downtime.
In order to upgrade a HDFS cluster without downtime, the cluster must be setup with HA. In order to upgrade a HDFS cluster without downtime, the cluster must be setup with HA.
</p> </p>
<p>
Note that rolling upgrade is supported only from Hadoop-2.4.0 onwards.
</p>
<subsection name="Upgrade without Downtime" id="UpgradeWithoutDowntime"> <subsection name="Upgrade without Downtime" id="UpgradeWithoutDowntime">
<p> <p>
In a HA cluster, there are two or more <em>NameNodes (NNs)</em>, many <em>DataNodes (DNs)</em>, In a HA cluster, there are two or more <em>NameNodes (NNs)</em>, many <em>DataNodes (DNs)</em>,