Yeah, my concern is that people should get Apache Spark from *Apache*, not from a vendor. It helps everyone use the latest features no matter where they are. In the Hadoop distro case, Hadoop made all this effort to have standard APIs (e.g. YARN), so it should be easy. But it is a problem if we're not packaging for the newest versions of some distros; I think we just fell behind at Hadoop 2.4.
Matei > On Mar 8, 2015, at 8:02 PM, Sean Owen <so...@cloudera.com> wrote: > > Yeah it's not much overhead, but here's an example of where it causes > a little issue. > > I like that reasoning. However, the released builds don't track the > later versions of Hadoop that vendors would be distributing -- there's > no Hadoop 2.6 build for example. CDH4 is here, but not the > far-more-used CDH5. HDP isn't present at all. The CDH4 build doesn't > actually work with many CDH4 versions. > > I agree with the goal of maximizing the reach of Spark, but I don't > know how much these builds advance that goal. > > Anyone can roll-their-own exactly-right build, and the docs and build > have been set up to make that as simple as can be expected. So these > aren't *required* to let me use latest Spark on distribution X. > > I had thought these existed to sorta support 'legacy' distributions, > like CDH4, and that build was justified as a > quasi-Hadoop-2.0.x-flavored build. But then I don't understand what > the MapR profiles are for. > > I think it's too much work to correctly, in parallel, maintain any > customizations necessary for any major distro, and it might be best to > do not at all than to do it incompletely. You could say it's also an > enabler for distros to vary in ways that require special > customization. > > Maybe there's a concern that, if lots of people consume Spark on > Hadoop, and most people consume Hadoop through distros, and distros > alone manage Spark distributions, then you de facto 'have to' go > through a distro instead of get bits from Spark? Different > conversation but I think this sort of effect does not end up being a > negative. > > Well anyway, I like the idea of seeing how far Hadoop-provided > releases can help. It might kill several birds with one stone. > > On Sun, Mar 8, 2015 at 11:07 PM, Matei Zaharia <matei.zaha...@gmail.com> > wrote: >> Our goal is to let people use the latest Apache release even if vendors fall >> behind or don't want to package everything, so that's why we put out >> releases for vendors' versions. It's fairly low overhead. >> >> Matei >> >>> On Mar 8, 2015, at 5:56 PM, Sean Owen <so...@cloudera.com> wrote: >>> >>> Ah. I misunderstood that Matei was referring to the Scala 2.11 tarball >>> at http://people.apache.org/~pwendell/spark-1.3.0-rc3/ and not the >>> Maven artifacts. >>> >>> Patrick I see you just commented on SPARK-5134 and will follow up >>> there. Sounds like this may accidentally not be a problem. >>> >>> On binary tarball releases, I wonder if anyone has an opinion on my >>> opinion that these shouldn't be distributed for specific Hadoop >>> *distributions* to begin with. (Won't repeat the argument here yet.) >>> That resolves this n x m explosion too. >>> >>> Vendors already provide their own distribution, yes, that's their job. >>> >>> >>> On Sun, Mar 8, 2015 at 9:42 PM, Krishna Sankar <ksanka...@gmail.com> wrote: >>>> Yep, otherwise this will become an N^2 problem - Scala versions X Hadoop >>>> Distributions X ... >>>> >>>> May be one option is to have a minimum basic set (which I know is what we >>>> are discussing) and move the rest to spark-packages.org. There the vendors >>>> can add the latest downloads - for example when 1.4 is released, HDP can >>>> build a release of HDP Spark 1.4 bundle. >>>> >>>> Cheers >>>> <k/> >>>> >>>> On Sun, Mar 8, 2015 at 2:11 PM, Patrick Wendell <pwend...@gmail.com> wrote: >>>>> >>>>> We probably want to revisit the way we do binaries in general for >>>>> 1.4+. IMO, something worth forking a separate thread for. >>>>> >>>>> I've been hesitating to add new binaries because people >>>>> (understandably) complain if you ever stop packaging older ones, but >>>>> on the other hand the ASF has complained that we have too many >>>>> binaries already and that we need to pare it down because of the large >>>>> volume of files. Doubling the number of binaries we produce for Scala >>>>> 2.11 seemed like it would be too much. >>>>> >>>>> One solution potentially is to actually package "Hadoop provided" >>>>> binaries and encourage users to use these by simply setting >>>>> HADOOP_HOME, or have instructions for specific distros. I've heard >>>>> that our existing packages don't work well on HDP for instance, since >>>>> there are some configuration quirks that differ from the upstream >>>>> Hadoop. >>>>> >>>>> If we cut down on the cross building for Hadoop versions, then it is >>>>> more tenable to cross build for Scala versions without exploding the >>>>> number of binaries. >>>>> >>>>> - Patrick >>>>> >>>>> On Sun, Mar 8, 2015 at 12:46 PM, Sean Owen <so...@cloudera.com> wrote: >>>>>> Yeah, interesting question of what is the better default for the >>>>>> single set of artifacts published to Maven. I think there's an >>>>>> argument for Hadoop 2 and perhaps Hive for the 2.10 build too. Pros >>>>>> and cons discussed more at >>>>>> >>>>>> https://issues.apache.org/jira/browse/SPARK-5134 >>>>>> https://github.com/apache/spark/pull/3917 >>>>>> >>>>>> On Sun, Mar 8, 2015 at 7:42 PM, Matei Zaharia <matei.zaha...@gmail.com> >>>>>> wrote: >>>>>>> +1 >>>>>>> >>>>>>> Tested it on Mac OS X. >>>>>>> >>>>>>> One small issue I noticed is that the Scala 2.11 build is using Hadoop >>>>>>> 1 without Hive, which is kind of weird because people will more likely >>>>>>> want >>>>>>> Hadoop 2 with Hive. So it would be good to publish a build for that >>>>>>> configuration instead. We can do it if we do a new RC, or it might be >>>>>>> that >>>>>>> binary builds may not need to be voted on (I forgot the details there). >>>>>>> >>>>>>> Matei >>>>> >>>>> --------------------------------------------------------------------- >>>>> To unsubscribe, e-mail: dev-unsubscr...@spark.apache.org >>>>> For additional commands, e-mail: dev-h...@spark.apache.org >>>>> >>>> >> --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@spark.apache.org For additional commands, e-mail: dev-h...@spark.apache.org
