Apache Ozone新版本无REST API？0.5.0版本API状态及第三方对接咨询

Great questions about Apache Ozone's API evolution and integration options—let's break this down clearly:

1. Was the alpha REST API removed in 0.5.0? #

Yes, the experimental REST API from early alpha versions was not carried forward into the 0.5.0 release. Ozone was still in heavy stabilization mode at this point, and the early REST interface was marked as purely experimental, meant for testing rather than production use. The team chose to prioritize core storage reliability and standardized integration paths over maintaining an unstable REST layer at this stage. It wasn't a permanent "removal" so much as a decision to not promote it to a supported feature.

2. Are there ways to communicate with Ozone without Java Client or Shell commands? #

Absolutely! The most accessible option is using Ozone's S3-compatible Object Store (OBS) interface. Ozone ships with an S3 gateway that implements the majority of Amazon S3's API specs, which means any application or tool that supports S3 can interact with Ozone directly—no custom Java client or shell scripts required.

For example, you can use curl to create a bucket:

curl -X PUT http://<your-ozone-s3g-host>:9878/my-new-bucket \
  -H "Authorization: AWS <your-access-key>:<your-secret-key>"

Or use standard S3 SDKs (like Python's boto3, Go's aws-sdk-go, etc.) by pointing the endpoint to your Ozone S3 gateway address.

3. How do third-party apps integrate with newer Ozone versions without a dedicated REST API? #

Beyond the S3 gateway, there are two primary supported paths:

• gRPC Interfaces: Ozone provides official gRPC service definitions (via .proto files) for its core services. You can generate client code in your preferred language (Java, Python, Go, etc.) from these definitions to build direct, high-performance integrations. This is ideal for applications that need low-latency access to Ozone's advanced features.
• Hadoop FileSystem Compatibility: Ozone implements the Hadoop FileSystem abstraction layer. If your application already works with HDFS, you can simply reconfigure it to use the ozone:// protocol instead of hdfs://, and it will interact with Ozone seamlessly. This is perfect for big data tools that rely on Hadoop's filesystem APIs.

If you have very specific needs, you could also build a custom integration on top of Ozone's internal protocols, but this is only recommended if you're familiar with Ozone's architecture—since the S3, gRPC, and Hadoop paths cover almost all common use cases.

内容的提问来源于stack exchange，提问作者dhalfageme