The problem was with the changes upstream. fetch upstream and a rebase resolved it and now the build is passing.
I also added a design doc and made the JIRA description a bit clearer (https://issues.apache.org/jira/browse/SPARK-24020) so I hope it will get merged soon. Thanks, Petar Sean Owen @ 1970-01-01 01:00 CET: > Also confused about this one as many builds succeed. One possible difference > is that this failure is in the Hive tests, so are you building and testing > with -Phive locally where it works? still does not explain the download > failure. It could be a mirror > problem, throttling, etc. But there again haven't spotted another failing > Hive test. > > On Wed, Jun 20, 2018 at 1:55 AM Petar Zecevic <petar.zece...@gmail.com> wrote: > > It's still dying. Back to this error (it used to be spark-2.2.0 before): > > java.io.IOException: Cannot run program "./bin/spark-submit" (in directory > "/tmp/test-spark/spark-2.1.2"): error=2, No such file or directory > So, a mirror is missing that Spark version... I don't understand why nobody > else has these errors and I get them every time without fail. > > Petar --------------------------------------------------------------------- To unsubscribe e-mail: dev-unsubscr...@spark.apache.org