ayo/ipex-llm
1
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1

[PPML] Modify the quickstart document of TPC-DS (#4441)

Browse source 
* Change command order
* [PPML]Modify the quickstart document of tpc-ds
* Remove blank
This commit is contained in:
Xiaohang Fu 2022-04-19 09:56:05 +08:00 committed by GitHub
parent 56aee11c82
commit 154f32ce97
1 changed files with 26 additions and 68 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

94
docs/readthedocs/source/doc/PPML/QuickStart/tpc-ds_with_sparksql_on_k8s.md
View file

@ -30,14 +30,35 @@ sbt "test:runMain com.databricks.spark.sql.perf.tpcds.GenTPCDSData -d <dsdgenDir
### Deploy PPML TPC-DS on Kubernetes
1. Pull docker image
1. Compile Kit
```
cd zoo-tutorials/tpcds-spark
sbt package
```
2. Create external tables
```
$SPARK_HOME/bin/spark-submit \
--class "createTables" \
--master <spark-master> \
--driver-memory 20G \
--executor-cores <executor-cores> \
--total-executor-cores <total-cores> \
--executor-memory 20G \
--jars spark-sql-perf/target/scala-2.12/spark-sql-perf_2.12-0.5.1-SNAPSHOT.jar \
target/scala-2.12/tpcds-benchmark_2.12-0.1.jar <dataDir> <dsdgenDir> <scaleFactor>
```
3. Pull docker image
```
sudo docker pull intelanalytics/bigdl-ppml-trusted-big-data-ml-python-graphene:2.1.0-SNAPSHOT
```
2. Prepare SGX keys, make sure keys and tpcds-spark can be accessed on each K8S node
3. Start a bigdl-ppml enabled Spark K8S client container with configured local IP, key, tpc-ds and kuberconfig path
4. Prepare SGX keys, make sure keys and tpcds-spark can be accessed on each K8S node
5. Start a bigdl-ppml enabled Spark K8S client container with configured local IP, key, tpc-ds and kuberconfig path
```
export ENCLAVE_KEY=/root/keys/enclave-key.pem
@ -75,13 +96,13 @@ sudo docker run -itd \
$DOCKER_IMAGE bash
```
4. Attach to the client container
6. Attach to the client container
```
sudo docker exec -it spark-local-k8s-client bash
```
5. Modify `spark-executor-template.yaml`, add path of `enclave-key`, `tpcds-spark` and `kuberconfig` on host
7. Modify `spark-executor-template.yaml`, add path of `enclave-key`, `tpcds-spark` and `kuberconfig` on host
```
apiVersion: v1
@ -110,69 +131,6 @@ spec:
path: /path/to/kuberconfig
```
6. Compile Kit
```
cd zoo-tutorials/tpcds-spark
sbt package
```
7. Create external tables
```
export TF_MKL_ALLOC_MAX_BYTES=10737418240 && \
export SPARK_LOCAL_IP=$LOCAL_IP && \
export HDFS_HOST=$hdfs_host_ip && \
export HDFS_PORT=$hdfs_port && \
export TPCDS_DIR=/ppml/trusted-big-data-ml/work/tpcds-spark \
export INPUT_DIR=$TPCDS_DIR/input \
export DSDGEN_DIR=tpcds-kit/tools \
export SCALE_FACTOR=1
/opt/jdk8/bin/java \
-cp '$TPCDS_DIR/spark-sql-perf/target/scala-2.12/spark-sql-perf_2.12-0.5.1-SNAPSHOT.jar:$TPCDS_DIR/target/scala-2.12/tpcds-benchmark_2.12-0.1.jar:$TPCH_DIR/input/*:$DSDGEN_DIR/*:/ppml/trusted-big-data-ml/work/spark-3.1.2/conf/:/ppml/trusted-big-data-ml/work/spark-3.1.2/jars/*' \
-Xmx10g \
-Dbigdl.mklNumThreads=1 \
org.apache.spark.deploy.SparkSubmit \
--master $RUNTIME_SPARK_MASTER \
--deploy-mode client \
--name spark-tpch-sgx \
--conf spark.driver.host=$LOCAL_IP \
--conf spark.driver.port=54321 \
--conf spark.driver.memory=10g \
--conf spark.driver.blockManager.port=10026 \
--conf spark.blockManager.port=10025 \
--conf spark.scheduler.maxRegisteredResourcesWaitingTime=5000000 \
--conf spark.worker.timeout=600 \
--conf spark.python.use.daemon=false \
--conf spark.python.worker.reuse=false \
--conf spark.network.timeout=10000000 \
--conf spark.starvation.timeout=250000 \
--conf spark.rpc.askTimeout=600 \
--conf spark.sql.autoBroadcastJoinThreshold=-1 \
--conf spark.io.compression.codec=lz4 \
--conf spark.sql.shuffle.partitions=8 \
--conf spark.speculation=false \
--conf spark.executor.heartbeatInterval=10000000 \
--conf spark.executor.instances=24 \
--executor-cores 8 \
--total-executor-cores 192 \
--executor-memory 16G \
--properties-file /ppml/trusted-big-data-ml/work/bigdl-2.1.0-SNAPSHOT/conf/spark-bigdl.conf \
--conf spark.kubernetes.authenticate.serviceAccountName=spark \
--conf spark.kubernetes.container.image=$RUNTIME_K8S_SPARK_IMAGE \
--conf spark.kubernetes.executor.podTemplateFile=/ppml/trusted-big-data-ml/spark-executor-template.yaml \
--conf spark.kubernetes.executor.deleteOnTermination=false \
--conf spark.kubernetes.executor.podNamePrefix=spark-tpcds-sgx \
--conf spark.kubernetes.sgx.enabled=true \
--conf spark.kubernetes.sgx.mem=32g \
--conf spark.kubernetes.sgx.jvm.mem=10g \
--class "createTables" \
--verbose \
--jars $TPCDS_DIR/spark-sql-perf/target/scala-2.12/spark-sql-perf_2.12-0.5.1-SNAPSHOT.jar \
$TPCDS_DIR/target/scala-2.12/tpcds-benchmark_2.12-0.1.jar \
$INPUT_DIR $DSDGEN_DIR $SCALE_FACTOR
```
8. Execute TPC-DS queries
Optional argument `QUERY` is the query number to run. Multiple query numbers should be separated by space, e.g. `1 2 3`. If no query number is specified, all 1-99 queries would be executed.