I'm using Hadoop 2.4.0 Hortonworks
I pretty new to Hadoop and am currently using Hue/Beeswax to run my standard queries which work fine, however, I now need to start creating basic shell scripts to do slightly more complex queries. Since I'm on a Win7 machine I've install Cygwin and using SSH to connect to a Hadoop server. When trying a simple test query I get the following error: -
hive -e "select actual_date from dm.date_dim limit 10"
AccessControlException: User blah cannot submit applications to queue root.default
At someone's suggestion I've tried changing the queue name first: -
set mapred.job.queuename=foo
Note, that if I run the same query but select all, i.e. so no MapReduce jobs are required, the query is successful: -
hive -e "select * from dm.date_dim limit 10"
My IT dept and other internal shell script developers are being very slow to respond to this (many days) so is there anything that you can suggest I should check or try. I was told that if changing the queue name had worked then I could've put that within my .hiverc file which I'm not sure where that is.
Is this a permissions thing? Should I be configuring something else?
I'd appreciate any help, but be aware I'm only started learning Linux very recently so you may need to dumb down any responses.
Your organization must have specified some queues to allocate the cluster resources to multiple teams. So any query other than 'select *' wont't run if you don't specify the job queue. So while running your hive queries just add this config and your queries will run:
hive -hiveconf mapred.job.queue.name=foo -e "select actual_date from dm.date_dim limit 10"