Hi, Then why not use the same way of specifying the queue name in the pig script and launch it. set mapreduce.job.queuename=queuename
~Abhishek > On Sep 13, 2016, at 1:48 PM, Bijoy Deb <[email protected]> wrote: > > Hi, > > When I submit a job using Pig command,I am able to submit it to the custom > queue using 'set mapreduce.job.queuename <name_of_queue>' command inside > Pig script.It seems the problem is when I specify the queuename in > <configuration> tag via Oozie shell action. > > We cannot use the Pig script directly and use shell action instead as we > have some custom operations such as logging,auditing,etc. that we want to > perform via the shell script during the course of execution of the Pig job > from within it. > > Thanks > > On Mon, Sep 12, 2016 at 1:44 AM, Peter Cseh <[email protected]> wrote: > >> Hi, >> >> Can you submit jobs into the specified queues outside Oozie? >> Is there any reason Pig Action >> <https://oozie.apache.org/docs/4.2.0/WorkflowFunctionalSpec.html# >> a3.2.3_Pig_Action> >> can't >> be used instead of Shell Action to submit Pig jobs? >> >> Thanks, >> Peter >> >> >> >> On Sat, Sep 10, 2016 at 7:37 PM, Bijoy Deb <[email protected]> >> wrote: >> >>> Hi, >>> I have configured 2 new queues 'launcher' and 'mapreduce' apart from >>> 'default' in Capacity scheduler in yarn-site. >>> Now, I am trying to submit a Pig job via *Oozie Shell action* into those >>> queues queues such that the oozie launcher job goes into 'launcher' and >> pig >>> job into 'mapreduce' queue.On submitting the oozie workflow I can see >> that >>> my launcher vjob successfully runs in 'launcher' queue,but the >>> Pig/mapreduce job is still running in default queue. >>> >>> Below is the snippet that I added to my workflow.xml for above: >>> >>> *<configuration>* >>> >>> *<property>* >>> >>> *<name>oozie.launcher.mapreduce.job.queuename</name>* >>> >>> *<value>launcher</value>* >>> >>> *</property>* >>> >>> *<property>* >>> >>> *<name>mapreduce.job.queuename</name>* >>> >>> *<value>mapreduce</value>* >>> *</property>* >>> *</configuration>* >>> >>> I also tried using 'mapred.job.queue.name' instead of >>> 'mapreduce.job.queuename' above and also by passing it as >>> -Dmapreduce.job.queuename=mapreduce while submitting the oozie job via >>> Oozie cli.But in every case the Pig/mapreduce job goes into 'default' >>> queue. >>> >>> Am I doing anything wrong here or is it that Oozie Shell action doesn't >>> support 'mapreduce.job.quenename specification? >>> >>> I am using MR2, Hadoop version 2.7.1, oozie version 4.2.0. >>> >>> Any help would be really appreciated. >>> >>> Thanks >>> >> >> >> >> -- >> Peter Cseh >> Software Engineer >> <http://www.cloudera.com> >>
