Dshackle Archive

Dshackle Archive is a tool that efficiently extracts blockchain data in JSON format and archives it into Avro files for scalable analysis using traditional Big Data tools. It supports Bitcoin and Ethereum-compatible blockchains and can operate in both batch mode for historical data and streaming mode for real-time data extraction.

Dshackle Archive copies JSON data from a blockchain to a plain files archive (i.e., it’s the Extraction of data, as in ETL). The archive is Avro files which contain blocks and transaction details fetched from blockchain nodes via their API. The Avro contain the JSON responses as is.

Features:

• Extracts data from Bitcoin and Ethereum compatible blockchains.

• Produces data in Avro format, where one file could be a range of blocks or a separate file per each block

• Archival can be scaled by using multiple blockchain nodes by using Dshackle Load Balancer.

• Runs in:

  • batch archive mode, when historical data is archived

  • or in a streaming mode when only fresh data is added to the archive

• Archive to:

  • Filesystem

  • S3

• Produces Notification to:

  • Filesystem

  • Apache Pulsar

The general idea it to use Dshackle Archive to copy data from a blockchain to a plain files, keeping the data structures as is, and then use traditional Big Data tools (Spark, Beam, etc.) to analyse data in a scalable way.

Note	It takes several weeks to archive the whole blockchain. It’s a very IO intensive operation, but can be scaled by using multiple nodes.

Note	It uses Dshackle protocol to connect to the blockchain, not a plain HTTP/JSON RPC. So a Dshackle instance is required to be running.

Usage

Command Line Options

Usage: dshackle-archive [OPTIONS] --blockchain <BLOCKCHAIN> --connection <HOST:PORT> <COMMAND>

Arguments:
  <COMMAND>  [possible values: stream, fix, verify, archive, compact]

Options:
  -b, --blockchain <BLOCKCHAIN>
          Blockchain
      --dryRun
          Do not modify the storage
  -c, --connection <HOST:PORT>
          Connection (host:port)
      --connection.notls
          Disable TLS
      --parallel <PARALLEL>
          How many requests to make in parallel. Range: 1..512. Default: 8 [default: 8]
      --notify.dir <NOTIFY_DIR>
          Write notifications as JSON line to the specified dir in a file <dshackle-archive-%STARTTIME.jsonl>
      --notify.pulsar.topic <PULSAR_TOPIC>
          Send notifications as JSON to the Pulsar to the specified topic (notify.pulsar.url must be specified)
      --notify.pulsar.url <PULSAR_URL>
          Send notifications as JSON to the Pulsar with specified URL (notify.pulsar.topic must be specified)
      --auth.aws.accessKey <ACCESS_KEY>
          AWS / S3 Access Key
      --auth.aws.secretKey <SECRET_KEY>
          AWS / S3 Secret Key
      --aws.endpoint <ENDPOINT>
          AWS / S3 endpoint url instead of the default one
      --aws.region <REGION>
          AWS / S3 region ID to use for requests
      --aws.s3.pathStyle
          Enable S3 Path Style access (default is false). Use this flag for a no-AWS service
      --aws.trustTls
          Trust any TLS certificate for AWS / S3 (default is false)
  -d, --dir <DIR>
          Target directory
      --continue
          Continue from the last file if set
      --tail <TAIL>
          Ensure that the latest T are archived (with `fix` command)
  -r, --range <RANGE>
          Blocks Range (`N...M`)
      --rangeChunk <RANGE_CHUNK>
          Range chunk size (default 1000)
  -h, --help
          Print help

Commands

• archive - the main operation, copies data from a blockchain to archive

• stream - append fresh blocks one by one to the archive

• compact - merge individual block files into larger range files

• fix - fix archive by making new archives for missing chunks

• verify - verify that archive files contains required data and delete incomplete files

Archive

The main operation, copies historical data from a blockchain to archive.

The data is copied in ranges, and with default range of 1000 blocks it produces two files per range. One for blocks in that range, and another one with all transactions in all blocks in that range.

See Archive Format.

Stream

Continuously append fresh blocks one by one to the archive. In addition to the copying, Dshackle archive can be configured to notify an external system about new blocks in the archive.

Note that when it’s in streaming mode the archives are writen in a per-block basis. I.e., each block comes with in a separate pair of two files. One for the block itself, and another one for all transactions in that block. To merge the individual files into larger ranges use compact command.

To notify an external system, there are two options:

• --notify.dir - write notifications as JSON line to the specified dir in a file <dshackle-archive-%STARTTIME.jsonl>

• --notify.pubsub - send notifications as JSON to the specified Google Pubsub topic

• --notify.pulsar.url + --notify.pulsar.topic - send notifications as JSON to the specified Apache Pulsar topic

See Notification format.

Compact

Merge individual block files into larger range files.

Fix

Fixes the archive by checking if there are any missing blocks, and if so, it creates new archives for the missing blocks.

Verify

Verify that archive files contains required data and delete incomplete/corrupted files. The a fix command is supposed to run to download missing blocks.

Warning

This command is destructive, it deletes files from the archive.

Archive Size

Dshackle Archive copies and stored data as JSON responses from blockchain nodes the resulting archive is much larger that the node database size, which keeps data in a compact format. It uses Snappy compression for Avro files, which give a good compression ratio, but still the resulting archive is large.

Average size of a 1000 blocks range (w/o expensive JSON such as stateDiff and trace):

• ~300Mb for Ethereum

• ~400Mb for Bitcoin

And the whole archive (w/o expensive JSON such as stateDiff and trace):

• ~2.5Tb for Ethereum

• ~1.9Tb for Bitcoin

Related projects:

• Avro structure and Java stubs: https://github.com/emeraldpay/dshackle-archive-avro

• Dshackle load balancer: https://github.com/emeraldpay/dshackle

FAQ

How to organize the data gathering process?

• First you need to archive the historical data, which may takes several week depending on how many and how fast nodes you have.

• After finishing the initial archive, you run in the Streaming mode which append new blocks to the archive as they are mined.

• Periodically (ex. once a day) you run Compaction to merge individual block files into larger range files.

• Also, periodically (ex. once a day) you run a pair of Verify and Fix commands to ensure the integrity of the archive.

What are supported blockchains?

Dshackle requires only compatibility onj JSON RPC level, so technically it can work with any blockchain that uses similar API. I.e., it’s compatible with all major blockchains, including Bitcoin, Ethereum, Binance Smart Chain, Polygon, etc.

What blockchain API it uses?

It uses Dshackle protocol to connect to the blockchain, not a plain HTTP/JSON RPC. So a Dshackle instance is required to be running.

Dshackle is a Load Balancer for Blockchain APIs, and it can route requests to multiple nodes, which scales up the archival throughput.

How does Dshackle Archive ensure the integrity and accuracy?

Dshackle provides two commands to ensure the integrity of the archive:

• first you run verify command, which checks the archive and deletes incomplete or corrupted files

• then you run fix command, which copies the data again for the blocks deleted in the previous step

You can schedule the execution of these commands to run periodically, e.g. once a day. To avoid scanning the whole archive every time, you can specify a range to check, e.g. --tail 1000. The option above specifies that it should verify/fix only the last 1000 blocks. I.e., it goes backward from the current head block.

Archive Format

For a complete descriptions, schema and libs to access Avro files please refer to https://github.com/emeraldpay/dshackle-archive-avro

Block

Fields common between different blockchains

• blockchainType - type of blockchain, as a definitions of what fields to expect. One of ETHEREUM or BITCOIN

• blockchainId - actual blockchain id (ETH, BTC, etc)

• archiveTimestamp - when the archive record was created. Milliseconds since epoch

• height - block height

• blockId - block hash

• timestamp - block timestamp. Milliseconds since epoch

• parentId - parent block hash

• json - JSON response for that block

Ethereum specific fields

• unclesCount - number of uncles for the current block

• uncle0Json - JSON for first uncle (eth_getUncleByBlockHashAndIndex(0))

• uncle1Json - JSON for second uncle (eth_getUncleByBlockHashAndIndex(1))

Bitcoin specific fields

• none

Transaction

Fields common between different blockchains

• blockchainType - type of blockchain, as a definitions of what fields to expect. One of ETHEREUM or BITCOIN

• blockchainId - blockchain id (ETH, BTC, etc)

• archiveTimestamp - when the archive record was created. Milliseconds since epoch

• height - block height

• blockId - block hash

• timestamp - block timestamp. Milliseconds since epoch

• index - index of the transaction in block

• txid - hash or transaction id of the transaction

• json - JSON response for that transaction

• raw - raw bytes of the transaction

Ethereum specific fields

• from - from address

• to - to address

• receiptJson - JSON response for eth_getTransactionReceipt

• traceJson - JSON response for trace_replayTransaction(trace)

• stateDiffJson - JSON response for trace_replayTransaction(stateDiff)

Bitcoin specific fields

• none

Notification format

{
  "version":"https://schema.emrld.io/dshackle-archive/notify",
  "ts":"2022-05-20T23:14:24.481327Z",
  "blockchain":"ETH",
  "type":"transactions",
  "run":"stream",
  "heightStart":14813875,
  "heightEnd":14813875,
  "location":"gs://my-bucket/blockchain-archive/eth/014000000/014813000/014813875.txes.avro"
}

Where

• version id of the current JSON format

• ts timestamp of the archive event

• blockchain blockchain

• type type of file (transactions or blocks)

• run mode in which the Dshackle Archive is run (archive, stream, copy or compact)

• heightStart and heightEnd range of blocks in the archived files

• location a URL to the archived file

Community

Development Chat

Join our Discord chat to discuss development and ask questions:

Commercial Support

Want to support the project, prioritize a specific feature, or get commercial help with using Dshackle in your project? Please contact [email protected] to discuss the possibility.

License

Copyright 2025 EmeraldPay

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.