Where does the classpath in spark-submit originate? Is compute-classpath.sh not the source?
As noted previously, the stable-ordering fix by me in compute-classpath.sh no longer seems to be effective either. Looks like some tracing of classpath assembly through the Spark command runner is required: https://github.com/apache/predictionio/blob/develop/tools/src/main/scala/org/apache/predictionio/tools/Runner.scala#L185 Unless someone with more knowledge of these internals could weigh-in… Donald? 😬😊 On Fri, Mar 9, 2018 at 15:44 Shane Johnson <[email protected]> wrote: > One additional item that you mentioned earlier is that we would need to > remove or skip the aws-java-sdk.jar that is already in the CLASSPATH. Do > you think this has impact? I did not write anything to skip or remove the > existing aws-java-sdk.jar. > > aws-java-sdk.jar is already in the CLASSPATH though, So, the script will >> need to skip or remove it first. > > > *Shane Johnson | LIFT IQ* > *Founder | CEO* > > *www.liftiq.com <http://www.liftiq.com/>* or *[email protected] > <[email protected]>* > mobile: (801) 360-3350 > LinkedIn <https://www.linkedin.com/in/shanewjohnson/> | Twitter > <https://twitter.com/SWaldenJ> | Facebook > <https://www.facebook.com/shane.johnson.71653> > > > > On Fri, Mar 9, 2018 at 4:41 PM, Shane Johnson <[email protected]> wrote: > >> Now that I am able to deploy I reset the buildpack to >> ...#debug-custom-dist and redeployed. Here is the build log...URL does >> point to the correct distribution with the edited compute-classpath.sh file. >> >> -----> JVM Common app detected >> >> -----> Installing JDK 1.8... done >> >> -----> PredictionIO app detected >> >> -----> Install core components >> >> + PredictionIO >> (https://s3-us-west-1.amazonaws.com/predictionio/0.12.0-incubating/apache-predictionio-0.12.0-incubating-bin.tar.gz) >> >> + Spark (spark-2.1.1-bin-hadoop2.7) >> >> -----> Install supplemental components >> >> + PostgreSQL (JDBC) >> >> + S3 HDFS (AWS SDK) >> >> + S3 HDFS (Hadoop-AWS) >> >> Writing default 'core-site.xml.erb' >> >> + local Maven repo from buildpack (contents) >> >> -----> Configure PredictionIO >> >> Writing default 'pio-env.sh' >> >> Writing default 'spark-defaults.conf.erb' >> >> + Maven repo from buildpack (build.sbt entry) >> >> Set-up environment via '.profile.d/' scripts >> >> -----> Install JVM (heroku/jvm-common) >> >> -----> PredictionIO engine >> >> Quietly logging. (Set `PIO_VERBOSE=true` for detailed build log.) >> >> [INFO] [Engine$] Using command >> '/tmp/build_67e7942abed821fccc839c9a79faf0eb/lift-iq-score-e92ed3de9212d04972e0e67e68b5407489e0c8d0/PredictionIO-dist/sbt/sbt' >> at >> /tmp/build_67e7942abed821fccc839c9a79faf0eb/lift-iq-score-e92ed3de9212d04972e0e67e68b5407489e0c8d0 >> to build. >> >> [INFO] [Engine$] If the path above is incorrect, this process will >> fail. >> >> [INFO] [Engine$] Uber JAR disabled. Making sure >> lib/pio-assembly-0.12.0-incubating.jar is absent. >> >> [INFO] [Engine$] Going to run: >> /tmp/build_67e7942abed821fccc839c9a79faf0eb/lift-iq-score-e92ed3de9212d04972e0e67e68b5407489e0c8d0/PredictionIO-dist/sbt/sbt >> package assemblyPackageDependency in >> /tmp/build_67e7942abed821fccc839c9a79faf0eb/lift-iq-score-e92ed3de9212d04972e0e67e68b5407489e0c8d0 >> >> [INFO] [Engine$] Compilation finished successfully. >> >> [INFO] [Engine$] Looking for an engine... >> >> [INFO] [Engine$] Found >> template-scala-parallel-liftscoring_2.11-0.1-SNAPSHOT.jar >> >> [INFO] [Engine$] Found >> template-scala-parallel-liftscoring-assembly-0.1-SNAPSHOT-deps.jar >> >> [INFO] [Engine$] Build finished successfully. >> >> [INFO] [Pio$] Your engine is ready for training. >> >> Using default Procfile for engine >> >> -----> Discovering process types >> >> Procfile declares types -> release, train, web >> >> -----> Compressing... >> >> Done: 376.7M >> >> >> The release log is below...I am not seeing the >> */app/PredictionIO-dist/lib/spark/aws-java-sdk.jar >> *show up at the beginning of the CLASSPATH, this is what we should see >> correct? I was also manipulating the compute-classpath.sh locally as well, >> I observed that adding a line right before echo "$CLASSPATH" was not >> changing what was in the logged spark-submit command as an FYI. This is >> what I was testing locally... >> >> >> CLASSPATH="*/Users/shanejohnson/Desktop/Apps/liftiq_platform/lift-s* >> *core/PredictionIO-dist/lib/spark/pio-data-s3-assembly-0.12.0-incubating.jar* >> :$CLASSPATH" >> echo "$CLASSPATH" >> >> I did not see any change in the spark-submit command by adding this when >> building and deploying locally. >> >> Release Log with new buildpack ..#debug-custom-dist >> >> Running train on release… >> >> Picked up JAVA_TOOL_OPTIONS: -Xmx12g -Dfile.encoding=UTF-8 >> >> [INFO] [Runner$] Submission command: >> /app/PredictionIO-dist/vendors/spark-hadoop/bin/spark-submit --driver-memory >> 13g --class org.apache.predictionio.workflow.CreateWorkflow --jars >> file:/app/PredictionIO-dist/lib/postgresql_jdbc.jar,file:/app/target/scala-2.11/template-scala-parallel-liftscoring-assembly-0.1-SNAPSHOT-deps.jar,file:/app/target/scala-2.11/template-scala-parallel-liftscoring_2.11-0.1-SNAPSHOT.jar,file:/app/PredictionIO-dist/lib/spark/._pio-data-hbase-assembly-0.12.0-incubating.jar,file:/app/PredictionIO-dist/lib/spark/pio-data-localfs-assembly-0.12.0-incubating.jar,file:/app/PredictionIO-dist/lib/spark/._pio-data-s3-assembly-0.12.0-incubating.jar,file:/app/PredictionIO-dist/lib/spark/._pio-data-localfs-assembly-0.12.0-incubating.jar,file:/app/PredictionIO-dist/lib/spark/pio-data-jdbc-assembly-0.12.0-incubating.jar,file:/app/PredictionIO-dist/lib/spark/pio-data-elasticsearch-assembly-0.12.0-incubating.jar,file:/app/PredictionIO-dist/lib/spark/._pio-data-elasticsearch-assembly-0.12.0-incubating.jar,file:/app/PredictionIO-dist/lib/spark/._pio-data-hdfs-assembly-0.12.0-incubating.jar,*file:/app/PredictionIO-dist/lib/spark/pio-data-s3-assembly-0.12.0-incubating.jar*,file:/app/PredictionIO-dist/lib/spark/pio-data-hbase-assembly-0.12.0-incubating.jar,file:/app/PredictionIO-dist/lib/spark/._pio-data-jdbc-assembly-0.12.0-incubating.jar,file:/app/PredictionIO-dist/lib/spark/hadoop-aws.jar,file:/app/PredictionIO-dist/lib/spark/pio-data-hdfs-assembly-0.12.0-incubating.jar,*file:/app/PredictionIO-dist/lib/spark/aws-java-sdk.jar* >> --files >> file:/app/PredictionIO-dist/conf/log4j.properties,file:/app/PredictionIO-dist/conf/core-site.xml >> --driver-class-path >> /app/PredictionIO-dist/conf:/app/PredictionIO-dist/conf:/app/PredictionIO-dist/lib/postgresql_jdbc.jar:/app/PredictionIO-dist/conf >> --driver-java-options -Dpio.log.dir=/app >> file:/app/PredictionIO-dist/lib/pio-assembly-0.12.0-incubating.jar >> --engine-id org.template.liftscoring.LiftScoringEngine --engine-version >> 0c35eebf403cf91fe77a64921d76aa1ca6411d20 --engine-variant >> file:/app/engine.json --verbosity 0 --json-extractor Both --env >> PIO_ENV_LOADED=1,PIO_EVENTSERVER_APP_NAME=classi,PIO_STORAGE_SOURCES_PGSQL_INDEX=enabled,PIO_S3_AWS_ACCESS_KEY_ID=AKIAJJX2S55QPCPZXGFQ,PIO_STORAGE_REPOSITORIES_METADATA_NAME=pio_meta,PIO_FS_BASEDIR=/app/.pio_store,PIO_STORAGE_SOURCES_ELASTICSEARCH_HOSTS=localhost,PIO_S3_BUCKET_NAME=lift-model-devmaster,PIO_EVENTSERVER_ACCESS_KEY=5954-20848-7512-17427-21660,PIO_HOME=/app/PredictionIO-dist,PIO_FS_ENGINESDIR=/app/.pio_store/engines,PIO_STORAGE_SOURCES_PGSQL_URL=jdbc:postgresql://ec2-52-70-46-243.compute-1.amazonaws.com:5432/dbvbo86hohutvb?sslmode=require,PIO_STORAGE_SOURCES_ELASTICSEARCH_TYPE=elasticsearch,PIO_STORAGE_REPOSITORIES_METADATA_SOURCE=PGSQL,PIO_SPARK_OPTS=--driver-memory >> 13g >> ,PIO_STORAGE_REPOSITORIES_MODELDATA_SOURCE=PGSQL,PIO_STORAGE_REPOSITORIES_EVENTDATA_NAME=pio_event,PIO_STORAGE_SOURCES_PGSQL_PASSWORD=p5c404ac780ab517d4ab249d7000809b51b4b987fdfb5c26e1bace511130337ac,PIO_STORAGE_SOURCES_ELASTICSEARCH_HOME=/app/PredictionIO-dist/vendors/elasticsearch,PIO_STORAGE_SOURCES_PGSQL_TYPE=jdbc,PIO_FS_TMPDIR=/app/.pio_store/tmp,PIO_STORAGE_SOURCES_PGSQL_USERNAME=ubefhv668b1s1m,PIO_STORAGE_REPOSITORIES_MODELDATA_NAME=pio_model,PIO_STORAGE_SOURCES_ELASTICSEARCH_SCHEMES=http,PIO_S3_AWS_SECRET_ACCESS_KEY=tQwL1PgYR0Y5MHG+qwVgEXNEcDcdlupaN2oO6JuR,PIO_TRAIN_SPARK_OPTS=--driver-memory >> 13g >> ,PIO_STORAGE_SOURCES_PGSQL_CONNECTIONS=8,PIO_STORAGE_REPOSITORIES_EVENTDATA_SOURCE=PGSQL,PIO_CONF_DIR=/app/PredictionIO-dist/conf,PIO_STORAGE_SOURCES_ELASTICSEARCH_PORTS=9200,PIO_STORAGE_SOURCES_PGSQL_PARTITIONS=4,PIO_S3_AWS_REGION=us-east-1 >> >> Picked up JAVA_TOOL_OPTIONS: -Xmx12g -Dfile.encoding=UTF-8 >> >> Picked up JAVA_TOOL_OPTIONS: -Xmx12g -Dfile.encoding=UTF-8 >> >> [INFO] [Engine] Extracting datasource params... >> >> [INFO] [Engine] Datasource params: (,DataSourceParams(Some(5))) >> >> [INFO] [Engine] Extracting preparator params... >> >> [WARN] [WorkflowUtils$] Non-empty parameters supplied to >> org.template.liftscoring.Preparator, but its constructor does not accept any >> arguments. Stubbing with empty parameters. >> >> [INFO] [Engine] Preparator params: (,Empty) >> >> >> >> >> >> *Shane Johnson | LIFT IQ* >> *Founder | CEO* >> >> *www.liftiq.com <http://www.liftiq.com/>* or *[email protected] >> <[email protected]>* >> mobile: (801) 360-3350 >> LinkedIn <https://www.linkedin.com/in/shanewjohnson/> | Twitter >> <https://twitter.com/SWaldenJ> | Facebook >> <https://www.facebook.com/shane.johnson.71653> >> >> >> >> On Fri, Mar 9, 2018 at 11:17 AM, Mars Hall <[email protected]> >> wrote: >> >>> I'm lost as to how such direct manipulation of CLASSPATH is not >>> appearing in the logged spark-submit command. >>> >>> What could cause this!? >>> >>> I just pushed a version of the buildpack which should help debug. >>> Assuming only a single buildpack is assigned to the app, here's how to set >>> it: >>> >>> heroku buildpacks:set >>> https://github.com/heroku/predictionio-buildpack#debug-custom-dist >>> >>> Then redeploy the engine an check the build log for the line: >>> >>> + PredictionIO ($URL) >>> >>> Please confirm that it is the URL of your custom PredictionIO dist. >>> >>> On Fri, Mar 9, 2018 at 2:47 PM, Shane Johnson <[email protected]> wrote: >>> >>>> Thanks Donald and Mars, >>>> >>>> I created a new distribution ( >>>> <https://s3-us-west-1.amazonaws.com/predictionio/0.12.0-incubating/apache-predictionio-0.12.0-incubating-bin.tar.gz> >>>> https://s3-us-west-1.amazonaws.com/predictionio/0.12.0-incubating/apache-predictionio-0.12.0-incubating-bin.tar.gz) >>>> with the added CLASSPATH code and pointed to the distribution with >>>> the PREDICTIONIO_DIST_URL variable within the engine app in Heroku. >>>> >>>> CLASSPATH="/app/PredictionIO-dist/lib/spark/aws-java-sdk.jar >>>> :$CLASSPATH" >>>> echo "$CLASSPATH" >>>> >>>> It didn't seem to force the aws-java-sdk to load first as I reviewed >>>> the release logs. Should the aws-java-sdk.jar show up as the first file >>>> within the --jars section when this is added CLASSPATH=" >>>> /app/PredictionIO-dist/lib/spark/aws-java-sdk.jar:$CLASSPATH". >>>> >>>> I'm still getting the NoSuchMethodError when the *aws-java-sdk.jar* loads >>>> after the *pio-data-s3-assembly-0.12.0-incubating.jar**. *Do you have >>>> other suggestions to try? I was also testing locally to change the order of >>>> the --jars but changes to the compute-classpath.sh didn't seem to change >>>> the order of the jars in the logs. >>>> >>>> Running train on release… >>>> >>>> Picked up JAVA_TOOL_OPTIONS: -Xmx12g -Dfile.encoding=UTF-8 >>>> >>>> [INFO] [Runner$] Submission command: >>>> /app/PredictionIO-dist/vendors/spark-hadoop/bin/spark-submit >>>> --driver-memory 13g --class >>>> org.apache.predictionio.workflow.CreateWorkflow --jars >>>> file:/app/PredictionIO-dist/lib/postgresql_jdbc.jar,file:/app/target/scala-2.11/template-scala-parallel-liftscoring-assembly-0.1-SNAPSHOT-deps.jar,file:/app/target/scala-2.11/template-scala-parallel-liftscoring_2.11-0.1-SNAPSHOT.jar,file:/app/PredictionIO-dist/lib/spark/pio-data-hdfs-assembly-0.12.0-incubating.jar,file:/app/PredictionIO-dist/lib/spark/pio-data-localfs-assembly-0.12.0-incubating.jar,file:/app/PredictionIO-dist/lib/spark/pio-data-elasticsearch-assembly-0.12.0-incubating.jar,file:/app/PredictionIO-dist/lib/spark/hadoop-aws.jar,file:/app/PredictionIO-dist/lib/spark/pio-data-hbase-assembly-0.12.0-incubating.jar,*file:/app/PredictionIO-dist/lib/spark/pio-data-s3-assembly-0.12.0-incubating.jar*,file:/app/PredictionIO-dist/lib/spark/pio-data-jdbc-assembly-0.12.0-incubating.jar,*file:/app/PredictionIO-dist/lib/spark/aws-java-sdk.jar* >>>> --files >>>> file:/app/PredictionIO-dist/conf/log4j.properties,file:/app/PredictionIO-dist/conf/core-site.xml >>>> --driver-class-path >>>> /app/PredictionIO-dist/conf:/app/PredictionIO-dist/conf:/app/PredictionIO-dist/lib/postgresql_jdbc.jar:/app/PredictionIO-dist/conf >>>> --driver-java-options -Dpio.log.dir=/app >>>> file:/app/PredictionIO-dist/lib/pio-assembly-0.12.0-incubating.jar >>>> --engine-id org.template.liftscoring.LiftScoringEngine --engine-version >>>> 0c35eebf403cf91fe77a64921d76aa1ca6411d20 --engine-variant >>>> file:/app/engine.json --verbosity 0 --json-extractor Both --env >>>> >>>> >>>> Error: >>>> >>>> Exception in thread "main" java.lang.NoSuchMethodError: >>>> com.amazonaws.services.s3.transfer.TransferManager.<init>(Lcom/amazonaws/services/s3/AmazonS3;Ljava/util/concurrent/ThreadPoolExecutor;)V >>>> >>>> at >>>> org.apache.hadoop.fs.s3a.S3AFileSystem.initialize(S3AFileSystem.java:287) >>>> >>>> at >>>> org.apache.hadoop.fs.FileSystem.createFileSystem(FileSystem.java:2669) >>>> >>>> at org.apache.hadoop.fs.FileSystem.access$200(FileSystem.java:94) >>>> >>>> >>>> >>>> >>>> >>>> *Shane Johnson | LIFT IQ* >>>> *Founder | CEO* >>>> >>>> *www.liftiq.com <http://www.liftiq.com/>* or *[email protected] >>>> <[email protected]>* >>>> mobile: (801) 360-3350 >>>> LinkedIn <https://www.linkedin.com/in/shanewjohnson/> | Twitter >>>> <https://twitter.com/SWaldenJ> | Facebook >>>> <https://www.facebook.com/shane.johnson.71653> >>>> >>>> >>>> >>>> On Wed, Mar 7, 2018 at 1:01 PM, Mars Hall <[email protected]> >>>> wrote: >>>> >>>>> Shane, >>>>> >>>>> On Wed, Mar 7, 2018 at 4:49 AM, Shane Johnson <[email protected]> >>>>> wrote: >>>>> >>>>>> >>>>>> Re: adding a line to ensure a jar is loaded first. Is this what you >>>>>> are referring to...(line at the bottom in red)? >>>>>> >>>>> >>>>> >>>>> I believe the code would need to look like this to effect the output >>>>> classpath as intended: >>>>> >>>>> >>>>>> CLASSPATH="/app/PredictionIO-dist/lib/spark/aws-java-sdk.jar >>>>>> :$CLASSPATH" >>>>>> echo "$CLASSPATH" >>>>>> >>>>> >>>>> >>>>> aws-java-sdk.jar is already in the CLASSPATH though, So, the script >>>>> will need to be skip or remove it first. >>>>> >>>>> -- >>>>> *Mars Hall >>>>> 415-818-7039 <(415)%20818-7039> >>>>> Customer Facing Architect >>>>> Salesforce Platform / Heroku >>>>> San Francisco, California >>>>> >>>> >>>> >>> >>> >>> -- >>> *Mars Hall >>> 415-818-7039 <(415)%20818-7039> >>> Customer Facing Architect >>> Salesforce Platform / Heroku >>> San Francisco, California >>> >> >> > -- *Mars Hall 415-818-7039 Customer Facing Architect Salesforce Platform / Heroku San Francisco, California
