Seems I find the solution here :

http://blog.cloudera.com/blog/2012/12/how-to-use-the-sharelib-in-apache-oozie/

and
http://grokbase.com/p/cloudera/cdh-user/125h1188yv/oozie-cant-find-hive-action-xsd



2013/3/13 Felix.徐 <[email protected]>

> Ok thanks!
>
> If oozie runs every task as a Mapreduce, does it mean that I have to
> install Hive on every machine if I want to execute Hive action?
>
> Oozie failed with NoClassDefFoundError:
> org/apache/hadoop/hive/cli/CliDriver currently.
>
>
> 2013/3/12 Alejandro Abdelnur <[email protected]>
>
>> Felix,
>>
>> On #1, you typically don't want to run user code (apps) in the oozie
>> server. Also, by using an MR job for the shell action we get automatic
>> load
>> distribution. If you need to run an action outside of the cluster, then
>> look at the SSH action.
>>
>> On #2, whatever you print in an action will be in the log of the launcher
>> MR job for the action in the only map task the job has. From the Oozie
>> WebUI you get can get to the JT job page, from there you must navigate to
>> the map task log.
>>
>> Thx
>>
>>
>> On Tue, Mar 12, 2013 at 12:45 AM, Felix.徐 <[email protected]> wrote:
>>
>> > I have 2 more questions:
>> > 1.Why is shell action performed as a mapreduce? Isn't it supposed to
>> run as
>> > a shell script on master node?
>> >
>> > 2.Is there any way to print some log within actions which can be shown
>> in
>> > the web console , current log seems meaningless to me ..
>> >
>> > Thanks!
>> >
>> >
>> > 2013/3/12 Felix.徐 <[email protected]>
>> >
>> > > Hi, I have 2 questions:
>> > > 1.How can I capture the output(a few lines) of a Hive query and pass
>> it
>> > to
>> > > the next action(e.g. a shell script) for further processing? The
>> > > capture-output seems only supports a properties file.
>> > >
>> > > 2.Is it possible to manually trigger the same workflow with different
>> > > parameters ?
>> > >
>> >
>>
>>
>>
>> --
>> Alejandro
>>
>
>

Reply via email to