Bump... Anyone out there running Zeppelin on ubuntu 14.04 using pyspark with Spark 1.3 and local[*] master?
On 18 August 2015 at 18:55, tab delimited <exception.bad...@gmail.com> wrote: > There's really nothing of interest about the code - any code introduced > with the %pyspark exhibits this issue. > > Do you have any examples of ubuntu 14.04 using local[*] master and spark > 1.3.0 that *don't *show this problem? If so perhaps we could compare > configs. > > Cheers, > /T > > On 15 August 2015 at 18:21, Felix Cheung <felixcheun...@hotmail.com> > wrote: > >> Hmm .. it could help if you could share the code/sample data to reproduce >> this. >> >> >> ------------------------------ >> Date: Sat, 15 Aug 2015 08:34:35 +0100 >> Subject: Re: pyspark "running" hang? >> From: exception.bad...@gmail.com >> To: users@zeppelin.incubator.apache.org >> >> >> There's nothing of interest in any of these three - anywhere else to look? >> >> We've independently reproduced this on several machines/environments, all >> Ubuntu 14.04. >> >> zeppelin-*.out >> zeppelin-*.log >> zeppelin-interpreter-spark-*.log >> >> /T >> >> On 11 August 2015 at 08:27, Felix Cheung <felixcheun...@hotmail.com> >> wrote: >> >> Could you check under the log directory for log files to see if there is >> any error? >> >> >> >> >> >> >> On Mon, Aug 10, 2015 at 1:08 PM -0700, "Exception Badger" < >> exception.bad...@gmail.com> wrote: >> >> Hi all, >> >> We've been using Zeppelin for a little while with CDH clusters and it's >> great. >> >> Recently a few of us have tried getting it working on local dev machines >> (ubuntu 14.04) without clusters, i.e. local[*] master and separately >> downloaded spark 1.3.0 referenced through spark.home. >> >> What we're seeing is pyspark notes hanging in the "running" state. >> >> Following a few suggestions on the web we've tried setting SPARK_HOME and >> PYTHONPATH explicitly both in the environment and also in the zeppelin >> config script. None of this seems to help. >> >> I've also tried building the branch and master and I see the same >> behaviour with both. >> >> It would be really good to get this working but we're kind of stumped. >> >> Any help appreciated! >> /T >> >> >> >> >