iamcaoxudong commented on a change in pull request #1700: HDFS-14963. Add HDFS
Client machine caching active namenode index mechanism.
URL: https://github.com/apache/hadoop/pull/1700#discussion_r348865504
##########
File path: hadoop-hdfs-project/hadoop-hdfs/src/main/resources/hdfs-default.xml
##########
@@ -1637,6 +1637,23 @@
</description>
</property>
+<property>
+ <name>dfs.client.failover.cache-active.enabled</name>
+ <value>true</value>
+ <description>
+ Whether caching active namenode index is enabled.
+ </description>
+</property>
+
+<property>
+ <name>dfs.client.failover.cache-active.dir</name>
+ <value>${java.io.tmpdir}</value>
Review comment:
In our environment, a HDFS client can dynamically load the core-site.xml and
hdfs-site.xml
which corresponding to the target hdfs uri (this is an enhancement we made
ourselves). In this case, there will be a lot of cache files ( As there are
many target HDFS clusters).
This kind of dynamic loading mechanism is still relatively common in a large
production environment. So I think it's better to put these cache files in a
less noticeable directory.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
