You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@zeppelin.apache.org by Jeff Zhang <zj...@gmail.com> on 2019/05/16 01:42:37 UTC
Re: “error: not found: value z” when using cluster mode
I tried branch-0.8, and it works for me, could you let me know what spark
version do you use ? and what is the stacktrace ?
Paul Brenner <pb...@placeiq.com> 于2019年5月16日周四 上午4:27写道：
> We are using 0.8.2 and can’t use z.show() with cluster mode. We get the
> following error:
>
> <console>:24: error: not found: value z z.show(
>
> This seems very similar to this issue which was marked resolved in 0.81:
>
> https://issues.apache.org/jira/browse/ZEPPELIN-3633
>
> Anyone familiar with this problem?
>
>
>
> PAUL BRENNER
>
> Head of Data Science
> pbrenner@placeiq.com | (217) 390-3033 | www.placeiq.com
> twitter @placeiq <https://twitter.com/PlaceIQ> linkedin /placeiq
> <https://www.linkedin.com/company/placeiq/>
> [image: Cannes in View]
> <http://pages.placeiq.com/FY19Q2-CannesMeeting_LP-Registration.html>
> [image: PlaceIQ: Oracle Data Cloud Premier Data Provider]
> <http://go.placeiq.com/WC01DPG0l0N3Gq000000e00>
>
>
>
>
>
>
--
Best Regards
Jeff Zhang
Re: “error: not found: value z” when using cluster mode
Posted by Jeff Zhang <zj...@gmail.com>.
Hi Paul,
I tried branch-0.8, and it works for me. The last commit
is a006affe46c0d4826aa2ddcb1493d743e2a33a8d and I use spark-2.3.3
https://user-images.githubusercontent.com/164491/58300715-4e53b600-7e16-11e9-9219-d6fc9618b98e.png
[image: image.png]
Paul Brenner <pb...@placeiq.com> 于2019年5月16日周四 上午9:59写道：
> Spark version 2.3.1 with scala 2.11.8
>
> No error shows up in the interpreter log, I only see errors in the
> zeppelin .log file which gives very limited info:
>
> WARN [2019-05-16 01:51:59,655] ({pool-2-thread-4}
> NotebookServer.java[afterStatusChange]:2316) - Job
> 20190515-171508_909555181 is finished, status: ERROR, exception: null,
> result: %text test: org.apache.spark.sql.DataFrame = [_c0: string, _c1:
> string ... 5 more fields]
> <console>:24: error: not found: value z
> z.show(
> ^
>
> I saw in the previous ticket that errors with zeppelin context and cluster
> deploy mode weren’t showing up in testing unless tests were run on a multi
> node cluster. I can confirm that we have a multinode cluster. Not sure how
> i would simulate a single node cluster to test if that makes a difference.
>
>
>
> PAUL BRENNER
>
> Head of Data Science
> pbrenner@placeiq.com | (217) 390-3033 | www.placeiq.com
> twitter @placeiq <https://twitter.com/PlaceIQ> linkedin /placeiq
> <https://www.linkedin.com/company/placeiq/>
> [image: Cannes in View]
> <http://pages.placeiq.com/FY19Q2-CannesMeeting_LP-Registration.html>
> [image: PlaceIQ: Oracle Data Cloud Premier Data Provider]
> <http://go.placeiq.com/WC01DPG0l0N3Gq000000e00>
>
>
>
>
>
> On May 15, 2019, 9:49 PM -0400, Jeff Zhang <zj...@gmail.com>, wrote:
>
> I tried branch-0.8, and it works for me, could you let me know what spark
> version do you use ? and what is the stacktrace ?
>
> Paul Brenner <pb...@placeiq.com> 于2019年5月16日周四 上午4:27写道：
>
> We are using 0.8.2 and can’t use z.show() with cluster mode. We get the
> following error:
>
> <console>:24: error: not found: value z z.show(
>
> This seems very similar to this issue which was marked resolved in 0.81:
>
> https://issues.apache.org/jira/browse/ZEPPELIN-3633
>
> Anyone familiar with this problem?
>
>
>
> PAUL BRENNER
>
> Head of Data Science
> pbrenner@placeiq.com | (217) 390-3033 | www.placeiq.com
> twitter @placeiq <https://twitter.com/PlaceIQ> linkedin /placeiq
> <https://www.linkedin.com/company/placeiq/>
> [image: Cannes in View]
> <http://pages.placeiq.com/FY19Q2-CannesMeeting_LP-Registration.html>
> [image: PlaceIQ: Oracle Data Cloud Premier Data Provider]
> <http://go.placeiq.com/WC01DPG0l0N3Gq000000e00>
>
>
>
>
>
>
>
> --
> Best Regards
>
> Jeff Zhang
>
>
--
Best Regards
Jeff Zhang
Re: “error: not found: valuez” when using cluster mode
Posted by Paul Brenner <pb...@placeiq.com>.
Spark version 2.3.1 with scala 2.11.8
No error shows up in the interpreter log, I only see errors in the zeppelin .log file which gives very limited info:
WARN [2019-05-16 01:51:59,655] ({pool-2-thread-4} NotebookServer.java[afterStatusChange]:2316) - Job 20190515-171508_909555181 is finished, status: ERROR, exception: null, result: %text test: org.apache.spark.sql.DataFrame = [_c0: string, _c1: string ... 5 more fields]
<console>:24: error: not found: value z
z.show(
^
I saw in the previous ticket that errors with zeppelin context and cluster deploy mode weren’t showing up in testing unless tests were run on a multi node cluster. I can confirm that we have a multinode cluster. Not sure how i would simulate a single node cluster to test if that makes a difference.
PAUL BRENNER
Head of Data Science
pbrenner@placeiq.com | (217) 390-3033 | www.placeiq.com
twitter @placeiq linkedin /placeiq
On May 15, 2019, 9:49 PM -0400, Jeff Zhang <zj...@gmail.com>, wrote:
> I tried branch-0.8, and it works for me, could you let me know what spark version do you use ? and what is the stacktrace ?
>
> > Paul Brenner <pb...@placeiq.com> 于2019年5月16日周四 上午4:27写道：
> > > We are using 0.8.2 and can’t use z.show() with cluster mode. We get the following error:
> > >
> > > <console>:24: error: not found: value z z.show(
> > >
> > > This seems very similar to this issue which was marked resolved in 0.81:
> > >
> > > https://issues.apache.org/jira/browse/ZEPPELIN-3633
> > >
> > > Anyone familiar with this problem?
> > >
> > >
> > >
> > > PAUL BRENNER
> > > Head of Data Science
> > > pbrenner@placeiq.com | (217) 390-3033 | www.placeiq.com
> > > twitter @placeiq linkedin /placeiq
> > >
> > >
> > >
> > >
> > >
>
>
> --
> Best Regards
>
> Jeff Zhang