On Tue, Mar 6, 2018 at 11:39 PM, Shane Johnson <[email protected]> wrote:
> > Do you know the version of hadoop-aws.jar and aws-java-sdk.jar that you > are using? > > I do not know what version is being used. Is this something that I can > specify or control? I am using the PredictionIO buildpack > https://github.com/heroku/predictionio-buildpack. I am *not* specifying > these in my build.sbt currently. > The versions are specified in the "bin/common/setup-runtime" script of the buildpack: https://github.com/heroku/predictionio-buildpack/blob/master/bin/common/setup-runtime#L50 Currently they are: - hadoop-aws-2.7.3 - aws-java-sdk-1.7.4 If you fork the buildpack, those download URLs (currently hosted in my S3 bucket) can be changed. > You are also right that you can modify the class path in > bin/compute-classpath.sh as a short term fix. The current order is > following the output of your target system's `ls`, so the order is not > guaranteed like you speak. Right before the last line (echo $CLASSPATH), > you can add a line to make sure that the JAR you want to be loaded first is > at the very beginning. > > I believe this would need to be something that is edited on master ( > https://github.com/apache/predictionio) as the buildpack leverages it vs. > a cloned version of the code that I can edit, am I thinking through that > correctly? I may need to circle back with Mars to see if there are any > other options to get this to work with Heroku. Is this something that can > be committed to master? The aws-java-sdk.jar is the one that we need to > load first. > If you build your own custom binary distribution of PredictionIO 0.12 and upload it to a publicly accessible URL, then the buildpack can be configured to use it by setting config var `PREDICTIONIO_DIST_URL` which is used by the same setup script as above: https://github.com/heroku/predictionio-buildpack/blob/master/bin/common/setup-runtime#L82 This is how I tested my pre-release builds on Heroku. -- *Mars Hall 415-818-7039 Customer Facing Architect Salesforce Platform / Heroku San Francisco, California
