[ https://issues.apache.org/jira/browse/IGNITE-1555?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14952771#comment-14952771 ]
Vladimir Ozerov commented on IGNITE-1555: ----------------------------------------- Guys, My main concern here is that the name should be easy to understand for end users. "ignite-hadoop" and "Ignite Hadoop Accelerator" are good names because users understand what they have in hands right away. "ignite-accelerator" raises questions what exactly we accelerate. "ignite-integration" is a better option, but it is inconsistent with the rest available Ignite integrations. Users simply would not understand why Hadoop and Spark are inside "integration" build, while Hibernate, Mesos, Kafka and a dozen of others are in "fabric" build. Can we leave the name "ignite-hadoop" as is and simply include Spark dependency into it? It could be a good compromise for both Ignite and BigTop products for now. > Combine ignite-hadoop and ignite-spark into signle assembly for downstream > consumption. > --------------------------------------------------------------------------------------- > > Key: IGNITE-1555 > URL: https://issues.apache.org/jira/browse/IGNITE-1555 > Project: Ignite > Issue Type: New Feature > Components: hadoop > Affects Versions: ignite-1.4 > Reporter: Konstantin Boudnik > Assignee: Konstantin Boudnik > Fix For: 1.5 > > Attachments: IGNITE-1555.patch, IGNITE-1555.patch, IGNITE-1555.patch > > > Let's combine all nice things that Ignite provides for Hadoop and Spark into > single assembly called ignite-accelerator. This will be advantageous for > downstream integrators like Bigtop, where all bits can be packaged together, > tested, and deployed with proper configurations and permissions to avoid > interoperability issues. A typical is when spark-shell starts an Ignite node > which crashes because user 'spark' isn't allowed to write into Ignite's > work-dir, etc. -- This message was sent by Atlassian JIRA (v6.3.4#6332)