My desktop is running Ubuntu 22.04.1 LTS, with JAVA_HOME pointing to
jdk1.8.0_341
I ran build with '-Pyarn -Pmesos -Pkubernetes' profiles [1] and with
$HOME/.m2 cleaned up.
Regards,
Mridul
[1] ARGS="-Pyarn -Pmesos -Pkubernetes"; ./build/mvn $ARGS clean &&
./build/mvn -DskipTests $ARGS package
Some users have observed issues like what you're describing related to the
job commit algorithm, which is controlled by configuration
property spark.hadoop.mapreduce.fileoutputcommitter.algorithm.version.
Hadoop's default value for this setting is 2. You can find a description of
the algorithms in
Could you provide your environment and test profile? Both community CIs
look fine to me.
GitHub Action:
https://github.com/apache/spark/actions?query=branch%3Abranch-3.3
Apple Silicon Jenkins Farm:
https://apache-spark.s3.fr-par.scw.cloud/BRANCH-3.3.html
Dongjoon.
On Fri, Oct 21, 2022 at 8:48
Hi,
I saw a couple of test failures I have not observed before:
a) FsHistoryProviderSuite - "SPARK-33146: don't let one bad rolling log
folder prevent loading other applications"
b) MesosClusterSchedulerSuite - "accept/decline offers with driver
constraints"
I ended up 'ignore''ing them to
