Leszek Wiesner 06712bb2e6 Merge branch 'giza' into giza-cli vor 3 Jahren
..
bin 8a858d797f storage-node-v2: Initialize CLI framework (oclif). vor 3 Jahren
scripts 06712bb2e6 Merge branch 'giza' into giza-cli vor 3 Jahren
src 3c7af9d630 storage-node-v2: Fix review comments. vor 3 Jahren
.eslintignore 5a74d847ef storage-node-v2: Add queries for query-node. vor 3 Jahren
.eslintrc.js 5cfe15b8ba storage-node-v2: Refactor the code. vor 3 Jahren
.gitignore a266ab9721 storage-node-v2: Add code comments. vor 3 Jahren
.prettierignore 97779b3def storage-node-v2: Improve ‘create storage bucket’ command. vor 3 Jahren
README.md 6f141b5074 storage-node-v2: Update README. vor 3 Jahren
package.json 3c7af9d630 storage-node-v2: Fix review comments. vor 3 Jahren
tsconfig.json 5485a9a2c3 Merge branch 'giza' into storage_node_v2_sync5 vor 3 Jahren

README.md

Colossus v2

Joystream storage subsystem.

oclif License

Description

The main responsibility of Colossus is handling media data for users. The data could be images, audio, or video files. Colossus receives uploads and saves files in the local folder, registers uploads in the blockchain, and later serves files to Argus nodes (distribution nodes). Colossus instances spread the data using peer-to-peer synchronization. Data management is blockchain-based, it relies on the concepts of buckets, bags, data objects. The full description of the blockchain smart contracts could be found here.

API

Colossus provides REST API for its clients and other Colossus instances. It's based on the OpenAPI Specification v3. Here is the complete spec (work in progress).

API endpoints:

  • files
    • get - get the data file by its ID
    • head - get the data file headers by its ID
    • post - upload a file
  • state

    • version - Colossus version and system environment
    • all data objects IDs
    • data objects IDs for bag
    • data statistics - total data folder size and data object number

      CLI

There is a command-line interface to manage Storage Working Group operations like create a bucket or change storage settings. Full description could be found below.

There are several groups of command:

  • leader - manages the Storage Working group in the blockchain. Requires leader privileges.
  • operator - Storage Provider group - it manages data object uploads and storage provider metadata(endpoint is the most important). Requires active Storage Working group membership.
  • dev - development support commands. Requires development blockchain setup with Alice account.
  • ungroupped - server and help commands. server starts Colossus and help shows the full command list.

Metadata

The storage provider should provide metadata for Colossus instance to be discoverable by other Colossus or Argus (distributor node) instances. At the very least an endpoint should be registered in the blockchain. For some complex scenarios, Colossus should provide its geolocation.

Metadata could be registered using operator:set-metadata command. A simple endpoint could be set using the --endpoint flag of the command. Complex metadata requires JSON file (example). JSON file format based on the protobuf format described here.

Data

Uploading

Colossus accepts files using its API. The data must be uploaded using POST http method with multipart/form-data. Simplified process (file uploading):

  • accepting the data upload in the temp folder
  • data hash & size verification
  • moving the data to the data folder
  • registering the data object as accepted in the blockchain

Synchronization

Several instances of Colossus should contain the data replica in order to provide some degree of reliability. When some Colossus instance receives the data upload it marks the related data object as accepted. Other instances that have the same obligations to store the data (they serve storage buckets assigned to the same bag) will eventually load this data object from the initial receiver (or some other node that already downloaded a new data object from the initial receiver) using REST API.

Distribution

The actual data distribution (serving to end-users) is done via Argus - the distributor node. It gets data from Colossus using the same get endpoint on a single data object basis.

Comments

Installation

# Ubuntu Linux

# Install packages required for installation
apt update
apt install git curl

# Clone the code repository
git clone https://github.com/Joystream/joystream
cd joystream

# Install volta
curl https://get.volta.sh | bash
bash

# Install project dependencies and build it
yarn
yarn workspace @joystream/types build
yarn workspace @joystream/metadata-protobuf build
yarn workspace storage-node-v2 build

# Verify installation
cd storage-node-v2
yarn storage-node version

Usage

$ yarn storage-node server --apiUrl ws://localhost:9944  -w 0 --accountUri //Alice -q localhost:8081 -o 3333 -d ~/uploads --sync

Prerequisites

  • accountURI or keyfile and password
  • workerId from the Storage working group that matches with the account above
  • Joystream node websocket endpoint URL
  • QueryNode URL
  • (optional) ElasticSearch URL
  • created directory for data uploading

CLI command

Full command description could be find below.

Docker

There is also an option to run Colossus as Docker container.

CLI Commands

storage-node dev:init

Initialize development environment. Sets Alice as storage working group leader.

USAGE
  $ storage-node dev:init

OPTIONS
  -h, --help                   show CLI help
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -m, --dev                    Use development mode
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/dev/init.ts

storage-node dev:multihash

Creates a multihash (blake3) for a file.

USAGE
  $ storage-node dev:multihash

OPTIONS
  -f, --file=file  (required) Path for a hashing file.
  -h, --help       show CLI help

See code: src/commands/dev/multihash.ts

storage-node dev:sync

Synchronizes the data - it fixes the differences between local data folder and worker ID obligations from the runtime.

USAGE
  $ storage-node dev:sync

OPTIONS
  -d, --uploads=uploads                              (required) Data uploading directory (absolute path).
  -h, --help                                         show CLI help

  -o, --dataSourceOperatorUrl=dataSourceOperatorUrl  [default: http://localhost:3333] Storage node url base (e.g.:
                                                     http://some.com:3333) to get data from.

  -p, --syncWorkersNumber=syncWorkersNumber          [default: 20] Sync workers number (max async operations in
                                                     progress).

  -q, --queryNodeEndpoint=queryNodeEndpoint          [default: http://localhost:8081/graphql] Query node endpoint (e.g.:
                                                     http://some.com:8081/graphql)

  -t, --syncWorkersTimeout=syncWorkersTimeout        [default: 30] Asset downloading timeout for the syncronization (in
                                                     minutes).

  -w, --workerId=workerId                            (required) Storage node operator worker ID.

See code: src/commands/dev/sync.ts

storage-node dev:upload

Upload data object (development mode only).

USAGE
  $ storage-node dev:upload

OPTIONS
  -c, --cid=cid                (required) Data object IPFS content ID.
  -h, --help                   show CLI help
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -m, --dev                    Use development mode
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -s, --size=size              (required) Data object size.
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/dev/upload.ts

storage-node dev:verify-bag-id

The command verifies bag id supported by the storage node. Requires chain connection.

USAGE
  $ storage-node dev:verify-bag-id

OPTIONS
  -h, --help
      show CLI help

  -i, --bagId=bagId
      (required) Bag ID. Format: {bag_type}:{sub_type}:{id}.
           - Bag types: 'static', 'dynamic'
           - Sub types: 'static:council', 'static:wg', 'dynamic:member', 'dynamic:channel'
           - Id:
             - absent for 'static:council'
             - working group name for 'static:wg'
             - integer for 'dynamic:member' and 'dynamic:channel'
           Examples:
           - static:council
           - static:wg:storage
           - dynamic:member:4

See code: src/commands/dev/verify-bag-id.ts

storage-node help [COMMAND]

display help for storage-node

USAGE
  $ storage-node help [COMMAND]

ARGUMENTS
  COMMAND  command to show help for

OPTIONS
  --all  see all commands in CLI

See code: @oclif/plugin-help

storage-node leader:cancel-invite

Cancel a storage bucket operator invite. Requires storage working group leader permissions.

USAGE
  $ storage-node leader:cancel-invite

OPTIONS
  -h, --help                   show CLI help
  -i, --bucketId=bucketId      (required) Storage bucket ID
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -m, --dev                    Use development mode
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/leader/cancel-invite.ts

storage-node leader:create-bucket

Create new storage bucket. Requires storage working group leader permissions.

USAGE
  $ storage-node leader:create-bucket

OPTIONS
  -a, --allow                  Accepts new bags
  -h, --help                   show CLI help
  -i, --invited=invited        Invited storage operator ID (storage WG worker ID)
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -m, --dev                    Use development mode
  -n, --number=number          Storage bucket max total objects number
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -s, --size=size              Storage bucket max total objects size
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/leader/create-bucket.ts

storage-node leader:delete-bucket

Delete a storage bucket. Requires storage working group leader permissions.

USAGE
  $ storage-node leader:delete-bucket

OPTIONS
  -h, --help                   show CLI help
  -i, --bucketId=bucketId      (required) Storage bucket ID
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -m, --dev                    Use development mode
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/leader/delete-bucket.ts

storage-node leader:invite-operator

Invite a storage bucket operator. Requires storage working group leader permissions.

USAGE
  $ storage-node leader:invite-operator

OPTIONS
  -h, --help                   show CLI help
  -i, --bucketId=bucketId      (required) Storage bucket ID
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -m, --dev                    Use development mode
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.
  -w, --operatorId=operatorId  (required) Storage bucket operator ID (storage group worker ID)

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/leader/invite-operator.ts

storage-node leader:remove-operator

Remove a storage bucket operator. Requires storage working group leader permissions.

USAGE
  $ storage-node leader:remove-operator

OPTIONS
  -h, --help                   show CLI help
  -i, --bucketId=bucketId      (required) Storage bucket ID
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -m, --dev                    Use development mode
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/leader/remove-operator.ts

storage-node leader:set-bucket-limits

Set VoucherObjectsSizeLimit and VoucherObjectsNumberLimit for the storage bucket.

USAGE
  $ storage-node leader:set-bucket-limits

OPTIONS
  -h, --help                   show CLI help
  -i, --bucketId=bucketId      (required) Storage bucket ID
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -m, --dev                    Use development mode
  -o, --objects=objects        (required) New 'voucher object number limit' value
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -s, --size=size              (required) New 'voucher object size limit' value
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/leader/set-bucket-limits.ts

storage-node leader:set-global-uploading-status

Set global uploading block. Requires storage working group leader permissions.

USAGE
  $ storage-node leader:set-global-uploading-status

OPTIONS
  -h, --help                   show CLI help
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -m, --dev                    Use development mode
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -s, --set=(on|off)           (required) Sets global uploading block (on/off).
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/leader/set-global-uploading-status.ts

storage-node leader:update-bag

Add/remove a storage bucket from a bag (adds by default).

USAGE
  $ storage-node leader:update-bag

OPTIONS
  -a, --add=add
      [default: ] ID of a bucket to add to bag

  -h, --help
      show CLI help

  -i, --bagId=bagId
      (required) Bag ID. Format: {bag_type}:{sub_type}:{id}.
           - Bag types: 'static', 'dynamic'
           - Sub types: 'static:council', 'static:wg', 'dynamic:member', 'dynamic:channel'
           - Id:
             - absent for 'static:council'
             - working group name for 'static:wg'
             - integer for 'dynamic:member' and 'dynamic:channel'
           Examples:
           - static:council
           - static:wg:storage
           - dynamic:member:4

  -k, --keyFile=keyFile
      Key file for the account. Mandatory in non-dev environment.

  -m, --dev
      Use development mode

  -p, --password=password
      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.

  -r, --remove=remove
      [default: ] ID of a bucket to remove from bag

  -u, --apiUrl=apiUrl
      [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.

  -y, --accountUri=accountUri
      Account URI (optional). Has a priority over the keyFile and password flags. Could be overriden by ACCOUNT_URI 
      environment variable.

See code: src/commands/leader/update-bag.ts

storage-node leader:update-bag-limit

Update StorageBucketsPerBagLimit variable in the Joystream node storage.

USAGE
  $ storage-node leader:update-bag-limit

OPTIONS
  -h, --help                   show CLI help
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -l, --limit=limit            (required) New StorageBucketsPerBagLimit value
  -m, --dev                    Use development mode
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/leader/update-bag-limit.ts

storage-node leader:update-blacklist

Add/remove a content ID from the blacklist (adds by default).

USAGE
  $ storage-node leader:update-blacklist

OPTIONS
  -a, --add=add                [default: ] Content ID to add
  -h, --help                   show CLI help
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -m, --dev                    Use development mode
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -r, --remove=remove          [default: ] Content ID to remove
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/leader/update-blacklist.ts

storage-node leader:update-bucket-status

Update storage bucket status (accepting new bags).

USAGE
  $ storage-node leader:update-bucket-status

OPTIONS
  -h, --help                   show CLI help
  -i, --bucketId=bucketId      (required) Storage bucket ID
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -m, --dev                    Use development mode
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -s, --set=(on|off)           (required) Sets 'accepting new bags' parameter for the bucket (on/off).
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/leader/update-bucket-status.ts

storage-node leader:update-data-fee

Update data size fee. Requires storage working group leader permissions.

USAGE
  $ storage-node leader:update-data-fee

OPTIONS
  -f, --fee=fee                (required) New data size fee
  -h, --help                   show CLI help
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -m, --dev                    Use development mode
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/leader/update-data-fee.ts

storage-node leader:update-dynamic-bag-policy

Update number of storage buckets used in the dynamic bag creation policy.

USAGE
  $ storage-node leader:update-dynamic-bag-policy

OPTIONS
  -h, --help                      show CLI help
  -k, --keyFile=keyFile           Key file for the account. Mandatory in non-dev environment.
  -m, --dev                       Use development mode
  -n, --number=number             (required) New storage buckets number
  -p, --password=password         Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -t, --bagType=(Channel|Member)  (required) Dynamic bag type (Channel, Member).
  -u, --apiUrl=apiUrl             [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.

  -y, --accountUri=accountUri     Account URI (optional). Has a priority over the keyFile and password flags. Could be
                                  overriden by ACCOUNT_URI environment variable.

See code: src/commands/leader/update-dynamic-bag-policy.ts

storage-node leader:update-voucher-limits

Update VoucherMaxObjectsSizeLimit and VoucherMaxObjectsNumberLimit for the Joystream node storage.

USAGE
  $ storage-node leader:update-voucher-limits

OPTIONS
  -h, --help                   show CLI help
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -m, --dev                    Use development mode
  -o, --objects=objects        (required) New 'max voucher object number limit' value
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -s, --size=size              (required) New 'max voucher object size limit' value
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/leader/update-voucher-limits.ts

storage-node operator:accept-invitation

Accept pending storage bucket invitation.

USAGE
  $ storage-node operator:accept-invitation

OPTIONS
  -h, --help                   show CLI help
  -i, --bucketId=bucketId      (required) Storage bucket ID
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -m, --dev                    Use development mode
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.
  -w, --workerId=workerId      (required) Storage operator worker ID

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/operator/accept-invitation.ts

storage-node operator:set-metadata

Set metadata for the storage bucket.

USAGE
  $ storage-node operator:set-metadata

OPTIONS
  -e, --endpoint=endpoint      Root distribution node endpoint
  -h, --help                   show CLI help
  -i, --bucketId=bucketId      (required) Storage bucket ID
  -j, --jsonFile=jsonFile      Path to JSON metadata file
  -k, --keyFile=keyFile        Key file for the account. Mandatory in non-dev environment.
  -m, --dev                    Use development mode
  -p, --password=password      Key file password (optional). Could be overriden by ACCOUNT_PWD environment variable.
  -u, --apiUrl=apiUrl          [default: ws://localhost:9944] Runtime API URL. Mandatory in non-dev environment.
  -w, --operatorId=operatorId  (required) Storage bucket operator ID (storage group worker ID)

  -y, --accountUri=accountUri  Account URI (optional). Has a priority over the keyFile and password flags. Could be
                               overriden by ACCOUNT_URI environment variable.

See code: src/commands/operator/set-metadata.ts

storage-node server

Starts the storage node server.

USAGE
  $ storage-node server

OPTIONS
  -d, --uploads=uploads                              (required) Data uploading directory (absolute path).

  -e, --elasticSearchEndpoint=elasticSearchEndpoint  Elasticsearch endpoint (e.g.: http://some.com:8081).
                                                     Log level could be set using the ELASTIC_LOG_LEVEL enviroment
                                                     variable.
                                                     Supported values: warn, error, debug, info. Default:debug

  -h, --help                                         show CLI help

  -i, --syncInterval=syncInterval                    [default: 1] Interval between synchronizations (in minutes)

  -k, --keyFile=keyFile                              Key file for the account. Mandatory in non-dev environment.

  -m, --dev                                          Use development mode

  -o, --port=port                                    (required) Server port.

  -p, --password=password                            Key file password (optional). Could be overriden by ACCOUNT_PWD
                                                     environment variable.

  -q, --queryNodeEndpoint=queryNodeEndpoint          (required) [default: http://localhost:8081/graphql] Query node
                                                     endpoint (e.g.: http://some.com:8081/graphql)

  -r, --syncWorkersNumber=syncWorkersNumber          [default: 20] Sync workers number (max async operations in
                                                     progress).

  -s, --sync                                         Enable data synchronization.

  -t, --syncWorkersTimeout=syncWorkersTimeout        [default: 30] Asset downloading timeout for the syncronization (in
                                                     minutes).

  -u, --apiUrl=apiUrl                                [default: ws://localhost:9944] Runtime API URL. Mandatory in
                                                     non-dev environment.

  -w, --worker=worker                                (required) Storage provider worker ID

  -y, --accountUri=accountUri                        Account URI (optional). Has a priority over the keyFile and
                                                     password flags. Could be overriden by ACCOUNT_URI environment
                                                     variable.

See code: src/commands/server.ts