README
¶
Overview
wasm is an EXPERIMENTAL simple example that loads and executes a wasm file function. greet.wasm, Cargo.toml and greet.rs were copied from the example provided by the wazero library: https://github.com/tetratelabs/wazero/blob/v1.0.0-pre.3/examples/allocation/rust/greet.go
Usage
To run this example in various runners, the following assumes:
OUTPUT=<path to output i.e. /tmp/example/output>
git clone https://github.com/apache/beam
BEAM_HOME=$(pwd)/beam
cd $BEAM_HOME/sdks
Direct Runner execution
To execute this example on the direct runner:
go run ./go/examples/wasm --output=$OUTPUT
Flink Portable Runner
The following describes how to execute this example on the flink portable runner.
REQUIREMENTS:
- Docker; MacOS users may consider alternative: https://github.com/abiosoft/colima
- Google Cloud Storage (https://cloud.google.com/storage) or S3 (https://aws.amazon.com/s3/) bucket; NOTE this example was only tested on Google Cloud Storage
0. Set OUTPUT to a Cloud storage bucket path
The example below shows Google Cloud Storage.
OUTPUT=gs://<my bucket>/greet
1. Find the latest flink runner version
cd $BEAM_HOME
./gradlew :runners:flink:properties --property flink_versions
Expected output should include the following, from which you acquire the latest flink runner version.
'flink_versions: 1.12,1.13,1.14,1.15'
2. Set to the latest flink runner version i.e. 1.15
FLINK_VERSION=1.15
3. In a separate terminal, start the flink runner (It should take a few minutes on the first execution)
cd $BEAM_HOME
./gradlew :runners:flink:$FLINK_VERSION:job-server:runShadow
Note the JobService host and port from the output, similar to:
INFO: JobService started on localhost:8099
4. Set the JOB_SERVICE variable from the aforementioned output
JOB_SERVICE=localhost:8099
5. Execute this example using the portable flink runner
cd $BEAM_HOME/sdks
go run ./go/examples/wasm --runner=universal --endpoint=$JOB_SERVICE --output=$OUTPUT --environment_config=apache/beam_go_sdk:latest
Dataflow Runner
The following describes how to execute this example on Dataflow.
REQUIREMENTS:
- Google Cloud Storage (https://cloud.google.com/storage) bucket
1. Set OUTPUT to a Cloud storage bucket path
The example below shows Google Cloud Storage.
OUTPUT=gs://<my bucket>/greet
2. Set additional variables
PROJECT=<project id>
REGION=<region>
STAGING=gs://<my bucket>/staging
NETWORK=<network>
SUBNETWORK=regions/$REGION/subnetworks/<subnetwork>
3. Execute this example using Dataflow
cd $BEAM_HOME/sdks
go run ./go/examples/wasm --runner=dataflow --output=$OUTPUT --environment_config=apache/beam_go_sdk:latest \
--project=$PROJECT --region=$REGION --network=$NETWORK --subnetwork=$SUBNETWORK --staging_location=$STAGING
Documentation
¶
Overview ¶
wasm is an EXPERIMENTAL simple example that loads and executes a wasm file function. greet.wasm, Cargo.toml and greet.rs were copied from the example provided by the wazero library: https://github.com/tetratelabs/wazero/blob/v1.0.0-pre.3/examples/allocation/rust/greet.go
New Concepts: 1. Load a wasm file compiled from: cargo build --release --target wasm32-unknown-unknown 2. Execute a wasm function within a DoFn
Source Files