Hello,
In our current project we have an oozie workflow which has pig actions where we
write entries to hive after data propagation. This requires hive properties to
be defined in the workflow and currently we’re referencing a copy of hive’s
configuration xml file through the <job-xml> element with all possible hive
properties. So what our oozie workflow xml looks like is something like this:
(pseudocode)
<workflow-app ...>
...
<action ...>
<pig>
<job-tracker/>
<name-node/>
<job-xml>/workflows/.../hive-conf.xml</>
<configuration>
<property/>
<property/>
<property/>
</configuration>
<script>/pigscript.pig</script>
<argument/>
<argument/>
<argument/>
</pig>
<ok to/>
<error to/>
</action>
...
</workflow-app>
The hive-conf.xml includes basically all of the possible hive properties
starting with authentication properties ending with file footer inclusion etc.
So the question would be that which configuration parameters does pig actually
use to communicate with hive and which parameters should we include in our
configuration element to make the workflow a bit cleaner?
Thanks in advance,
Marko
