[
https://issues.apache.org/jira/browse/FLUME-2286?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14158207#comment-14158207
]
Hari Shreedharan commented on FLUME-2286:
-----------------------------------------
Actually the issue is that not just minicluster, even Hadoop and Hbase depend
on really old guava versions. If we upgrade it in Flume, there will be issues
in actual prod environment where we pull Hadoop and hbase classpaths into ours.
We'd end up having two very different versions in our classpath.
One workaround would be to shade guava, to keep us independent of Hadoop and
Hbase.
> Upgrade Guava version to 14.0.1
> -------------------------------
>
> Key: FLUME-2286
> URL: https://issues.apache.org/jira/browse/FLUME-2286
> Project: Flume
> Issue Type: Wish
> Affects Versions: v1.5.0
> Reporter: Joshua Hyde
> Priority: Minor
> Labels: patch
> Attachments: FLUME-2286.patch
>
>
> I wrote a custom Sink as a Flume plugin that uses code that leverages the
> {{CacheBuilder}} class provided by Guava. The problem was that my code uses
> Guava 14.0.1, and I ended up with this error:
> {noformat}
> 31 Dec 2013 11:11:45,071 ERROR
> [SinkRunner-PollingRunner-DefaultSinkProcessor]
> (my.package.CustomSink.process:104) - Failed to process events.
> java.lang.NoSuchMethodError:
> com.google.common.cache.CacheBuilder.maximumSize(J)Lcom/google/common/cache/CacheBuilder;
> at my.package.CustomSink.process(CustomSink.java:90)
> at
> org.apache.flume.sink.DefaultSinkProcessor.process(DefaultSinkProcessor.java:68)
> at org.apache.flume.SinkRunner$PollingRunner.run(SinkRunner.java:147)
> at java.lang.Thread.run(Thread.java:744)
> {noformat}
> Thankfully, I have direct control over the code that uses Guava 14.0.1 and
> can download it to the version used by Flume (10.0.1 at time of writing
> this), but it'd be nice if I didn't have to downgrade my code base for this.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
