You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@spark.apache.org by shane knapp <sk...@berkeley.edu> on 2018/11/06 00:30:06 UTC
need assistance debugging a strange build failure...
i'm really close (for real: really close!) on the ubuntu port... but one
build has been a thorn in my side and i was wondering if i could get some
extra eyes on this as i grind through the remaining few pieces of my own
personal system dependency hell. :)
the job in question is:
https://amplab.cs.berkeley.edu/jenkins/job/spark-master-test-sbt-hadoop-2.7-ubuntu-testing
it's identical to the regular spark-master-test-sbt-hadoop-2.7 job, except
i'm building against a newer version of java (1.8.0_171 vs 1.8.0_60).
the centos job always passes on every worker.
the ubuntu job fails on every ubuntu worker during the scala unidoc
generation w/the following error:
"""
[error]
/home/jenkins/workspace/spark-master-test-sbt-hadoop-2.7-ubuntu-testing/core/src/main/scala/org/apache/spark/serializer/GenericAvroSerializer.scala:123:
value createDatumWriter is not a member of
org.apache.avro.generic.GenericData
[error] writerCache.getOrElseUpdate(schema,
GenericData.get.createDatumWriter(schema))
[error] ^
[info] No documentation generated with unsuccessful compiler run
[error] one error found
"""
an example job w/this failure is here:
https://amplab.cs.berkeley.edu/jenkins/job/spark-master-test-sbt-hadoop-2.7-ubuntu-testing/30/consoleFull
thoughts? am i missing something obvious? i've checked and there are no
avro system packages installed on any of the workers (centos or ubuntu).
thanks in advance,
shane
--
Shane Knapp
UC Berkeley EECS Research / RISELab Staff Technical Lead
https://rise.cs.berkeley.edu
Re: need assistance debugging a strange build failure...
Posted by shane knapp <sk...@berkeley.edu>.
btw, this is a compilation error in the SBT build that only shows up on the
ubuntu workers.
On Mon, Nov 5, 2018 at 5:07 PM shane knapp <sk...@berkeley.edu> wrote:
> the maven build is quite happy:
>
> https://amplab.cs.berkeley.edu/jenkins/job/spark-master-test-maven-hadoop-2.7-ubuntu-testing/
>
> i've wiped all the cached deps from .ivy2 on one ubuntu worker, pinned the
> build to that (research-jenkins-worker-08) and re-launched the sbt build.
> let's see if it fails in a new, spectacular way (or just the same as
> before).
>
> but again: why is this failing on the ubuntu nodes, vs centos? if it's a
> poisoned cache, i will be sad, but that's easy to fix. if it's a bad dep
> in the pom, then... ¯\_(ツ)_/¯
>
> On Mon, Nov 5, 2018 at 5:00 PM Wenchen Fan <cl...@gmail.com> wrote:
>
>> Have you tried Maven instead of SBT? This looks like a Java dependency
>> problem, e.g. a wrong version of Avro is picked.
>>
>> On Tue, Nov 6, 2018 at 8:30 AM shane knapp <sk...@berkeley.edu> wrote:
>>
>>> i'm really close (for real: really close!) on the ubuntu port... but
>>> one build has been a thorn in my side and i was wondering if i could get
>>> some extra eyes on this as i grind through the remaining few pieces of my
>>> own personal system dependency hell. :)
>>>
>>> the job in question is:
>>>
>>> https://amplab.cs.berkeley.edu/jenkins/job/spark-master-test-sbt-hadoop-2.7-ubuntu-testing
>>>
>>> it's identical to the regular spark-master-test-sbt-hadoop-2.7 job,
>>> except i'm building against a newer version of java (1.8.0_171 vs
>>> 1.8.0_60).
>>>
>>> the centos job always passes on every worker.
>>>
>>> the ubuntu job fails on every ubuntu worker during the scala unidoc
>>> generation w/the following error:
>>>
>>> """
>>> [error]
>>> /home/jenkins/workspace/spark-master-test-sbt-hadoop-2.7-ubuntu-testing/core/src/main/scala/org/apache/spark/serializer/GenericAvroSerializer.scala:123:
>>> value createDatumWriter is not a member of
>>> org.apache.avro.generic.GenericData
>>> [error] writerCache.getOrElseUpdate(schema,
>>> GenericData.get.createDatumWriter(schema))
>>> [error] ^
>>> [info] No documentation generated with unsuccessful compiler run
>>> [error] one error found
>>>
>>> """
>>> an example job w/this failure is here:
>>>
>>> https://amplab.cs.berkeley.edu/jenkins/job/spark-master-test-sbt-hadoop-2.7-ubuntu-testing/30/consoleFull
>>>
>>> thoughts? am i missing something obvious? i've checked and there are
>>> no avro system packages installed on any of the workers (centos or ubuntu).
>>>
>>> thanks in advance,
>>>
>>> shane
>>> --
>>> Shane Knapp
>>> UC Berkeley EECS Research / RISELab Staff Technical Lead
>>> https://rise.cs.berkeley.edu
>>>
>>
>
> --
> Shane Knapp
> UC Berkeley EECS Research / RISELab Staff Technical Lead
> https://rise.cs.berkeley.edu
>
--
Shane Knapp
UC Berkeley EECS Research / RISELab Staff Technical Lead
https://rise.cs.berkeley.edu
Re: need assistance debugging a strange build failure...
Posted by shane knapp <sk...@berkeley.edu>.
the maven build is quite happy:
https://amplab.cs.berkeley.edu/jenkins/job/spark-master-test-maven-hadoop-2.7-ubuntu-testing/
i've wiped all the cached deps from .ivy2 on one ubuntu worker, pinned the
build to that (research-jenkins-worker-08) and re-launched the sbt build.
let's see if it fails in a new, spectacular way (or just the same as
before).
but again: why is this failing on the ubuntu nodes, vs centos? if it's a
poisoned cache, i will be sad, but that's easy to fix. if it's a bad dep
in the pom, then... ¯\_(ツ)_/¯
On Mon, Nov 5, 2018 at 5:00 PM Wenchen Fan <cl...@gmail.com> wrote:
> Have you tried Maven instead of SBT? This looks like a Java dependency
> problem, e.g. a wrong version of Avro is picked.
>
> On Tue, Nov 6, 2018 at 8:30 AM shane knapp <sk...@berkeley.edu> wrote:
>
>> i'm really close (for real: really close!) on the ubuntu port... but one
>> build has been a thorn in my side and i was wondering if i could get some
>> extra eyes on this as i grind through the remaining few pieces of my own
>> personal system dependency hell. :)
>>
>> the job in question is:
>>
>> https://amplab.cs.berkeley.edu/jenkins/job/spark-master-test-sbt-hadoop-2.7-ubuntu-testing
>>
>> it's identical to the regular spark-master-test-sbt-hadoop-2.7 job,
>> except i'm building against a newer version of java (1.8.0_171 vs
>> 1.8.0_60).
>>
>> the centos job always passes on every worker.
>>
>> the ubuntu job fails on every ubuntu worker during the scala unidoc
>> generation w/the following error:
>>
>> """
>> [error]
>> /home/jenkins/workspace/spark-master-test-sbt-hadoop-2.7-ubuntu-testing/core/src/main/scala/org/apache/spark/serializer/GenericAvroSerializer.scala:123:
>> value createDatumWriter is not a member of
>> org.apache.avro.generic.GenericData
>> [error] writerCache.getOrElseUpdate(schema,
>> GenericData.get.createDatumWriter(schema))
>> [error] ^
>> [info] No documentation generated with unsuccessful compiler run
>> [error] one error found
>>
>> """
>> an example job w/this failure is here:
>>
>> https://amplab.cs.berkeley.edu/jenkins/job/spark-master-test-sbt-hadoop-2.7-ubuntu-testing/30/consoleFull
>>
>> thoughts? am i missing something obvious? i've checked and there are no
>> avro system packages installed on any of the workers (centos or ubuntu).
>>
>> thanks in advance,
>>
>> shane
>> --
>> Shane Knapp
>> UC Berkeley EECS Research / RISELab Staff Technical Lead
>> https://rise.cs.berkeley.edu
>>
>
--
Shane Knapp
UC Berkeley EECS Research / RISELab Staff Technical Lead
https://rise.cs.berkeley.edu
Re: need assistance debugging a strange build failure...
Posted by Wenchen Fan <cl...@gmail.com>.
Have you tried Maven instead of SBT? This looks like a Java dependency
problem, e.g. a wrong version of Avro is picked.
On Tue, Nov 6, 2018 at 8:30 AM shane knapp <sk...@berkeley.edu> wrote:
> i'm really close (for real: really close!) on the ubuntu port... but one
> build has been a thorn in my side and i was wondering if i could get some
> extra eyes on this as i grind through the remaining few pieces of my own
> personal system dependency hell. :)
>
> the job in question is:
>
> https://amplab.cs.berkeley.edu/jenkins/job/spark-master-test-sbt-hadoop-2.7-ubuntu-testing
>
> it's identical to the regular spark-master-test-sbt-hadoop-2.7 job, except
> i'm building against a newer version of java (1.8.0_171 vs 1.8.0_60).
>
> the centos job always passes on every worker.
>
> the ubuntu job fails on every ubuntu worker during the scala unidoc
> generation w/the following error:
>
> """
> [error]
> /home/jenkins/workspace/spark-master-test-sbt-hadoop-2.7-ubuntu-testing/core/src/main/scala/org/apache/spark/serializer/GenericAvroSerializer.scala:123:
> value createDatumWriter is not a member of
> org.apache.avro.generic.GenericData
> [error] writerCache.getOrElseUpdate(schema,
> GenericData.get.createDatumWriter(schema))
> [error] ^
> [info] No documentation generated with unsuccessful compiler run
> [error] one error found
>
> """
> an example job w/this failure is here:
>
> https://amplab.cs.berkeley.edu/jenkins/job/spark-master-test-sbt-hadoop-2.7-ubuntu-testing/30/consoleFull
>
> thoughts? am i missing something obvious? i've checked and there are no
> avro system packages installed on any of the workers (centos or ubuntu).
>
> thanks in advance,
>
> shane
> --
> Shane Knapp
> UC Berkeley EECS Research / RISELab Staff Technical Lead
> https://rise.cs.berkeley.edu
>