Oozie to drill down to the hadoop job id created by Pig
--------------------------------------------------------
Key: OOZIE-30
URL: https://issues.apache.org/jira/browse/OOZIE-30
Project: Apache Oozie (Incubating)
Issue Type: Improvement
Reporter: Mohammad Kamrul Islam
At present, Oozie supports to drill down till to MR job and the Launcher Mapper
job (for pig/Java action). However, Oozie doesn't have a drill down option of
going to the hadoop jobs created by pig scripts.
Users want to see/access to those job ids through WS API/CLI/UI.
In existing release, Oozie collect those ids in a hacky way (after parsing the
pig log file). In the latest Pig, pig allows to expose the stats that contains
the pig-created hadoop jobs ids (most possibly). We could utilize this new
approach to get the ids. Also , at present, we don't store that in databases
that could be sent as part of query. In future release. we could add an entry
to store these ids in DB table. In addition, we should allow to return those as
part of job query and shows those in the UI as well.
The Question is:
Should we expose this as job ids only? Or we can expose those as additional
external URLs that could be used through UI to jump to the corresponding JT
page.
Comments?
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
