Hi,
I have two data-sets.
<datasets>
<dataset name="input" frequency="${coord:hours(1)}"
initial-instance="${fileStartTs}" timezone="${cronTZ}">
<uri-template>${dataRoot}/year=${YEAR}/month=${MONTH}/day=${DAY}/hour=${HOUR}/min=00</uri-template>
<done-flag></done-flag>
</dataset>
<dataset name="output" frequency="${coord:hours(1)}"
initial-instance="${fileStartTs}" timezone="${cronTZ}">
<uri-template>${dataRoot}/year=${YEAR}/month=${MONTH}/day=${DAY}/hour=${HOUR}/min=00</uri-template>
</dataset>
</datasets>
My requirement is like :
1) data-set "input", I just need folder to be present, don't want to wait
for _SUCCESS file to be present.
2) data-set "output", which is my output location of the workflow, I want a
_SUCCESS file to be created. But after the completion of job, I don't see a
_SUCCESS file. I have tried without giving <done-flag/> and specifying
<done-flag>_SUCCESS</done-flag> as well. No luck. But [1] says, if you
don't specify anything, hadoop creates _SUCCESS file in the output
directory.
[1]
http://incubator.apache.org/oozie/docs/3.1.3-incubating/docs/CoordinatorFunctionalSpec.html#a5.1._Synchronous_Datasets
--
Thanks & Regards,
Bhargav
