Cloned from launchpad blueprint https://blueprints.launchpad.net/sahara/+spec/edp-datasource-placeholders. Description: Common use case: user wants to run EDP job two times. Now the only way to do that with the same data sources is to erase result of the first run before running job the second time. Allowing to have random part in URL will allow to use output with random suffix. Specification URL (additional information): None
Feature is implemented in upstream L.
The placeholders in datasources (%RANDSTR(<len>)%, %JOB_EXEC_ID%) are correctly replaced according the spec for both HDFS and Swift data sources. There is only a minor issue: when <len> is really big, for example 2^32+1 but also smaller values, something goes wrong and the job stays in PREPARE state. But this is really a corner case and it comes from a deliberate choice of the user that defines the data source, who can easily avoid the problem. Therefore I don't think this is a blocker for the general availability of the feature. The issue has been captured in the following upstream bug: https://bugs.launchpad.net/sahara/+bug/1560890 Verified on an up-to-date RHEL 7.2, with openstack-sahara-api-3.0.1-4.el7ost.noarch openstack-sahara-common-3.0.1-4.el7ost.noarch openstack-sahara-engine-3.0.1-4.el7ost.noarch python-sahara-3.0.1-4.el7ost.noarch python-sahara-tests-3.0.1-4.el7ost.noarch
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHEA-2016-0603.html