Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

exp/services/ledgerexporter: Add guide for contributors/developers #5340

Merged
merged 11 commits into from
Jun 27, 2024
65 changes: 65 additions & 0 deletions exp/services/ledgerexporter/DEVELOPER_GUIDE.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,65 @@

# Ledger Exporter Developer Guide
The ledger exporter is a tool to export Stellar network transaction data to cloud storage in a way that is easy to access.

## Prerequisites
This document assumes that you have installed and can run the ledger exporter, and that you have familiarity with its CLI and configuration. If not, please refer to the [Installation Guide](./README.md).

## Goal
The goal of the ledger exporter is to build an easy-to-use tool to export Stellar network ledger data to a configurable remote data store, such as cloud blob storage.
- Use cloud storage optimally
- Minimize network usage to export
- Make it easy and fast to search for a specific ledger or ledger range

## Architecture
To achieve its goals, the ledger exporter uses the following architecture, which consists of the 3 main components:
- Captive-core to extract raw transaction metadata from the Stellar Network.
- Export manager to bundles and organizes the ledgers to get them ready for export.
- The cloud storage plugin writes to the cloud storage. This is specific to the type of cloud storage, GCS in this case.
urvisavla marked this conversation as resolved.
Show resolved Hide resolved


![ledgerexporter-architecture](./architecture.png)


## Data Format
- Ledger exporter uses a compact and efficient data format called [XDR](https://developers.stellar.org/docs/learn/encyclopedia/data-format/xdr) (External Data Representation), which is a compact binary format. A Stellar Captive Core instance emits data in this format and the data structure is referred to as `LedgerCloseMeta`. The exporter bundles multiple `LedgerCloseMeta`'s into a single object using a custom XDR structure called `LedgerCloseMetaBatch` which is defined in [Stellar-exporter.x](https://github.com/stellar/go/blob/master/xdr/Stellar-exporter.x).

- The metadata for the same batch is also stored alongside each exported object. Supported metadata is defined in [metadata.go](https://github.com/stellar/go/blob/master/support/datastore/metadata.go).

- Objects are compressed before uploading using the [zstd](http://facebook.github.io/zstd/) (zstandard) compression algorithm to optimize network usage and storage needs.

## Data Storage
- An example implementation of `DataStore` for GCS, Google Cloud Storage. This plugin is located in the [support](https://github.com/stellar/go/tree/master/support/datastore) package.
- The ledger exporter currently implements the interface only for Google Cloud Storage (GCS). The [GCS plugin](https://github.com/stellar/go/blob/master/support/datastore/gcs_datastore.go) uses GCS-specific behaviors like conditional puts, automatic retry, metadata, and CRC checksum.

## Build, Run and Test using Docker
The Dockerfile contains all the necessary dependencies (e.g., Stellar-core) required to run the ledger exporter.

- Build: To build the Docker container, use the provided [Makefile](./Makefile). Simply run make `make docker-build` to build a new container after making any changes.

- Run: For instructions on running the Docker container, refer to the [Installation Guide](./README.md).

- Test: To test the Docker container, refer to the [docker-test](./Makefile) command for an example of how to use the [GCS emulator](https://github.com/fsouza/fake-gcs-server) for local testing.

## Adding support for a new storage type
Support for different data storage types are encapsulated as 'plugins', which are implementation of `DataStore` interface in a go package. To add a data storage plugin based on a new storage type (e.g. AWS S3), follow these steps:

- A data storage plugin must implement the [DataStore](https://github.com/stellar/go/blob/master/support/datastore/datastore.go) interface.
- Add support for new datastore-specific features. Implement any datastore-specific custom logic. Different datastores have different ways of handling
- race conditions
- automatic retries
- metadata storage, etc.
- Add the new datastore to the factory function [NewDataStore](https://github.com/stellar/go/blob/master/support/datastore/datastore.go).
- Add a [config](./config.example.toml) section for the new storage type. This may include configurations like destination, authentication information etc.
- An emulator such as a GCS emulator [fake-gcs-server](https://github.com/fsouza/fake-gcs-server) can be used for testing without connecting to real cloud storage.

### Design DOs and DONTs
- Multiple exporters should be able to run in parallel without the need for explicit locking or synchronization.
- Exporters when restarted do not have any memory of prior operation and rely on the already exported data as much as possible to decide where to resume.

## Using exported data
The exported data in storage can be used in the ETL pipeline to gather analytics and reporting. To write a tool that consumes exported data you can use Stellar ingestion library's `ledgerbackend` package. This package includes a ledger backend called [BufferedStorageBackend](https://github.com/stellar/go/blob/master/ingest/ledgerbackend/buffered_storage_backend.go),
which imports data from the storage and validates it. For more details, refer to the ledgerbackend [documentation](https://github.com/stellar/go/tree/master/ingest/ledgerbackend).

## Contributing
For information on how to contribute, please refer to our [Contribution Guidelines](https://github.com/stellar/go/blob/master/CONTRIBUTING.md).
Binary file added exp/services/ledgerexporter/architecture.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading