[
https://issues.apache.org/jira/browse/FLUME-1137?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13275146#comment-13275146
]
Juhani Connolly edited comment on FLUME-1137 at 5/16/12 5:23 AM:
-----------------------------------------------------------------
Looks like we need it for serializing data as writables.
We'd have to go with another serialization method if we chose not to add in
hadoop jars
this is kind of relevent: FLUME-776
was (Author: juhanic):
Looks like we need it for serializing data as writables.
We'd have to go with another serialization method if we chose not to add in
hadoop jars
> Recoverable Memory channel requires Hadoop classes in classpath
> ---------------------------------------------------------------
>
> Key: FLUME-1137
> URL: https://issues.apache.org/jira/browse/FLUME-1137
> Project: Flume
> Issue Type: Bug
> Components: Sinks+Sources
> Affects Versions: v1.2.0
> Reporter: Hari Shreedharan
> Fix For: v1.2.0
>
>
> RecoverableMemoryChannel needs Hadoop IO classes in classpath. On a machine
> where Hadoop isn't on Classpath, Class.forName throws NoClassDefFoundError,
> which is not caught by the factory class, causing the agent startup to fail.
> It is not exactly a good idea to expect every machine on which this channel
> will be used to have Hadoop installed. We currently depend on Hadoop/HDFS
> only where we are actually writing to HDFS. Even the JDBC channel has an
> embedded DB, thus not requiring the DB to be available in classpath.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
