0chain

        # Züs TestNet Setup with Docker Containers

• Züs Overview
• Changelog
• Initial Setup
  • Host Machine Network Setup
  • Directory Setup for Miners & Sharders
  • Setup Network
  • Building the Nodes
  • Configuring the Nodes
  • Starting the Nodes
  • Check Chain Status
  • Restarting the Nodes
  • Cleanup
• Run 0chain on ec2 / vm / bare metal
• Run 0chain on ec2 / vm / bare metal over https
• Development
  • Installing msgp
  • Dependencies for local compilation
  • Debugging
  • Unit tests
  • Creating The Magic Block
  • Initial states
  • Integration tests
    • Architecture
    • Running Integration Tests
    • Running Standard Tests
    • Running complex scenario suites
    • Running Blobber Tests
    • Adding new Tests
    • Supported Conductor Commands
    • Creating Custom Conductor Commands
  • Benchmarks
  • Swagger documentation

Züs is a high-performance cloud on a fast blockchain offering privacy and configurable uptime. It is an alternative to traditional cloud S3 and has shown better performance on a test network due to its parallel data architecture. The technology uses erasure code to distribute the data between data and parity servers. Züs storage is configurable to provide flexibility for IT managers to design for desired security and uptime, and can design a hybrid or a multi-cloud architecture with a few clicks using Blimp's workflow, and can change redundancy and providers on the fly.

For instance, the user can start with 10 data and 5 parity providers and select where they are located globally, and later decide to add a provider on-the-fly to increase resilience, performance, or switch to a lower cost provider.

Users can also add their own servers to the network to operate in a hybrid cloud architecture. Such flexibility allows the user to improve their regulatory, content distribution, and security requirements with a true multi-cloud architecture. Users can also construct a private cloud with all of their own servers rented across the globe to have a better content distribution, highly available network, higher performance, and lower cost.

The QoS protocol is time-based where the blockchain challenges a provider on a file that the provider must respond within a certain time based on its size to pass. This forces the provider to have a good server and data center performance to earn rewards and income.

The privacy protocol from Züs is unique where a user can easily share their encrypted data with their business partners, friends, and family through a proxy key sharing protocol, where the key is given to the providers, and they re-encrypt the data using the proxy key so that only the recipient can decrypt it with their private key.

Züs has ecosystem apps to encourage traditional storage consumption such as Blimp, a S3 server and cloud migration platform, and Vult, a personal cloud app to store encrypted data and share privately with friends and family, and Chalk, a high-performance story-telling storage solution for NFT artists.

Other apps are Bolt, a wallet that is very secure with air-gapped 2FA split-key protocol to prevent hacks from compromising your digital assets, and it enables you to stake and earn from the storage providers; Atlus, a blockchain explorer and Chimney, which allows anyone to join the network and earn using their server or by just renting one, with no prior knowledge required.

CHANGELOG.md

Docker and Go must be installed to run the testnet containers. Get Docker from here and Go from here.

./macos_network.sh

Run powershell as administrator

./windows_network.ps1

Run the following script

./wsl_ubuntu_network_iptables.sh

In the git/0chain run the following command

./docker.local/bin/init.setup.sh

Set up a network called testnet0 for each of these node containers to talk to each other.

Note: The config file should be providing the IP address of the nodes as per the IP addresses in this network.

./docker.local/bin/setup.network.sh

1. Open 5 terminal tabs. Use the first one for building the containers by being in git/0chain directory. Use the next 3 for 3 miners and be in the respective miner directories created above in docker.local. Use the 5th terminal and be in the sharder1 directory.

   1.1) First build the base containers, zchain_build_base and zchain_run_base

   ./docker.local/bin/build.base.sh
   

2. Build mocks from the Makefile in the repo, from git/0chain directory run:

    make build-mocks 
   

   Note: Mocks have to be built once in the beginning. Building mocks require mockery and brew which can be installed from here.

3. Building the miners and sharders. From the git/0chain directory use

   3.1) To build the miner containers

   ./docker.local/bin/build.miners.sh
   

   3.2) To build the sharder containers

   ./docker.local/bin/build.sharders.sh
   

   3.3) Syncing time (the host and the containers are being offset by a few seconds that throws validation errors as we accept transactions that are within 5 seconds of creation). This step is needed periodically when you see the validation error.

   ./docker.local/bin/sync_clock.sh
   

1. Use ./docker.local/config/0chain.yaml to configure the blockchain properties. The default options are set up for running the blockchain fast in development.

1.1) If you want the logs to appear on the console - change logging.console from false to true

1.2) If you want the debug statements in the logs to appear - change logging.level from "info" to "debug"

1.3) If you want to change the block size, set the value of server_chain.block.size

1.4) If you want to adjust the network relay time, set the value of network.relay_time

1.5) If you want to turn off fees adjust server_chain.smart_contract.miner from true to false

Note: Remove sharder72 and miner75 from docker.local/config/b0snode2_keys.txt and docker.local/config/b0mnode5_keys.txt respectively if you are joining to local network.

1. Starting the nodes. On each of the miner terminals use the commands (note the .. at the beginning. This is because, these commands are run from within the docker.local/<miner/sharder|i> directories and the bin is one level above relative to these directories)

Start sharder first because miners need the genesis magic block. On the sharder terminal, use

../bin/start.b0sharder.sh

Wait till the cassandra is started and the sharder is ready to listen to requests.

On the respective miner terminal, use

../bin/start.b0miner.sh

1. Ensure the port mapping is all correct:

docker ps

This should display a few containers and should include containers with images miner1_miner, miner2_miner and miner3_miner, and they should have the ports mapped like "0.0.0.0:7071->7071/tcp"

2. Confirming the servers are up and running. From a browser, visit

• http://localhost:7071/_diagnostics

• http://localhost:7072/_diagnostics

• http://localhost:7073/_diagnostics

to see the status of the miners.

Similarly, following links can be used to see the status of the sharders

• http://localhost:7171/_diagnostics

• http://localhost:7172/_diagnostics

• http://localhost:7173/_diagnostics

3. Connecting to redis servers running within the containers (you are within the appropriate miner directories)

Default redis (used for clients and state):

../bin/run.miner.sh redis redis-cli

Redis used for transactions:

../bin/run.miner.sh redis_txns redis-cli

4. Connecting to cassandra used in the sharder (you are within the appropriate sharder directories)

../bin/run.sharder.sh cassandra cqlsh

To reflect a change in config files 0chain.yaml and sc.yaml, just restart the miner or sharder to take the new configuration. If you're doing a code change locally or pulling updates from GitHub, you need to build.

git pull
docker.local/bin/build.base.sh && docker.local/bin/build.sharders.sh && docker.local/bin/build.miners.sh

For existing code and if you have tried running once, make sure there are no previous files and processes.

docker stop $(docker ps -a -q)
docker.local/bin/clean.sh
docker.local/bin/init.setup.sh
docker.local/bin/sync_clock.sh

Then go to individual miner/sharder:

../bin/start.b0sharder.sh (start sharders first!)
../bin/start.b0miner.sh

1. If you want to restart the blockchain from the beginning

./docker.local/bin/clean.sh

This cleans up the directories within docker.local/miner* and docker.local/sharder*

Note: this script can take a while if the blockchain generated a lot of blocks as the script deletes the databases and also all the blocks that are stored by the sharders. Since each block is stored as a separate file, deleting thousands of such files will take some time.

2. If you want to get rid of old unused docker resources:

docker system prune

Library by herumi for working with BLS threshold signatures requires this flag turned on:

setsebool -P selinuxuser_execheap 1

If you are curious about the reasons for this, this thread sheds some light on the topic:

https://github.com/herumi/xbyak/issues/9

The following is no longer required as the schema is automatically loaded.

Start the sharder service that also brings up the cassandra service. To run commands on cassandra, use the following command

../bin/run.sharder.sh cassandra cqlsh

1. To create zerochain keyspace, do the following

../bin/run.sharder.sh cassandra cqlsh -f /0chain/sql/zerochain_keyspace.sql

2. To create the tables, do the following

../bin/run.sharder.sh cassandra cqlsh -k zerochain -f /0chain/sql/txn_summary.sql

3. When you want to truncate existing data (use caution), do the following

../bin/run.sharder.sh cassandra cqlsh -k zerochain -f /0chain/sql/truncate_tables.sql

Run the following command to install the msgp tool:

make install-msgp

We are using msgp to encode/decode data that store in MPT, it is unnecessary to touch it unless there are data struct changes or new type of data structs need to store in MPT.

When we need to add a new data struct to MPT, for example:

//go:generate msgp -io=false -tests=false -v
package main

type Foo struct {
	Name string
}

Note:

1. msgp does not support system type alias, so please do not use datastore.Key in MPT data struct, it is an alias of system type string.
2. The //go:generate msgp -io=false ... works on file level, i.e, we only need to define it once a file, so please check if it is already defined before adding.

Then run the following command from the project root to generate methods for serialization.

make msgp

A new file will then be generated as {file}_gen.go in the same dir where the data struct is defined.

You need to install rocksdb and herumi/bls, refer to docker.local/build.base/Dockerfile.build_base for necessary steps.

For local compilation it should be enough of go build from a submodule folder, e.g.

cd code/go/0chain.net/miner
go build

You can pass tag development if you want to simulate n2n delays. And you also need tag bn256 to build the same code as in production:

go build -tags "bn256 development"

./0chain_dev_deployment.sh

If you want to run a debug 0chain build you can follow the details contained in the 0chain/local folder.

Only one miner and one sharder can be run on any single machine, so you will need at least three machines to for a working 0chain.

The logs of the nodes are stored in log directory (/0chain/log on the container and docker.local/miner|sharder[n]/log in the host). The 0chain.log contains all the logs related to the protocol and the n2n.log contains all the node to node communication logs. The typical issues that need to be debugged is errors in the log, why certain things have not happened which requires reviewing the timestamp of a sequence of events in the network. Here is an example set of commands to do some debugging.

Find errors in all the miner nodes (from git/0chain)

grep ERROR docker.local/miner*/log/0chain.log

This gives a set of errors in the log. Say an error indicates a problem for a specific block, say abc, then

grep abc docker.local/miner*/log/0chain.log

gives all the logs related to block 'abc'

To get the start time of all the rounds

grep 'starting round' docker.local/miner*/log/0chain.log

This gives the start timestamps that can be used to correlate the events and their timings.

0chain unit tests verify the behaviour of individual parts of the program. A config for the base docker image can be provided on run to execute general unit tests.

Navigate to 0chain folder and run the script to build base docker image for unit testing :

cd 0chain
./docker.local/bin/build.base.sh

The base image includes all the dependencies required to test the 0chain code.

Now run the script containing unit tests .

./docker.local/bin/unit_test.sh

OR to run the unit tests without the mocks,

./docker.local/bin/unit_test.sh --no-mocks 

The list of packages is optional, and if provided runs only the tests from those packages. The command for running unit tests with specific packages.

./docker.local/bin/unit_test.sh [<packages>]

Unit testing happens over a series of steps one after the other.

This FROMstep does the required preparation and specifies the underlying OS architecture to use the build image. Here we are using the base image created in the build phase.

The SRC_DIR variable is a reference to a filepath which contains the code from your pull request. Here /0chain directory is specified as it is the one which was cloned.

GO111MODULE is an environment variable that can be set when using go for changing how Go imports packages. It was introduced to help ensure a smooth transition to the module system.

GO111MODULE=on will force using Go modules even if the project is in your GOPATH. Requires go.mod to work.

Note: The default behavior in Go 1.16 is now GO111MODULE=on

This step copies the code from the source path to the destination path.

The RUN command is an image build step which allows installing of application and packages requited for testing while thego mod download downloads the specific module versions you've specified in the go.modfile.

This step runs the gozstd package and provides write permissions to the directory. gozstd which is a go wrapper for zstd (library) provides Go bindings for the libzstd C library. The make clean is run in the last to clean up the code and remove all the compiled object files from the source code

This step defines the working directory for running unit tests which is (0chain/code/go/0chain.net/).For all the running general unit tests their code coverage will be defined in the terminal like this

ok      0chain.net/chaincore/block      0.128s  coverage: 98.9% of statements

The above output shows 98.9% of code statements was covered with tests.

First build the magic block image.

./docker.local/bin/build.magic_block.sh

Next, set the configuration file. To do this edit the docker.local/build.magicBlock/docker-compose.yml file. On line 13 is a flag "--config_file" set it to the magic block configuration file you want to use.

To create the magic block.

./docker.local/bin/create.magic_block.sh

The magic block and the dkg summary json files will appear in the docker.local/config under the name given in the configuration file.

The magic_block_file setting in the 0chain.yaml file needs to be updated with the new name of the magic block created.

Update the miner config file, so it is set to the new dkg summaries. To do this edit the docker.local/build.miner/b0docker-compose.yml file. On line 55 is a flag "--dkg_file" set it to the dkg summary files created with the magic block.

The balance for the various nodes is set up in a initial_state.yaml file. This file is a list of node ids and token amounts.

The initial state yaml file is entered as a command line argument when running a sharder or miner, falling that the 0chain.yaml network.inital_states entry is used to find the initial state file.

An example, that can be used with the preset ids, can be found at 0chain/docker.local/config/initial_state.yaml`

Benchmark 0chain smart-contract endpoints.

Runs testing.Benchmark on each 0chain endpoint. The blockchain database used in these tests is constructed from the parameters in the benchmark.yaml. file. Smartcontracts do not (or should not) access tha chain so a populated MPT database is enough to give a realistic benchmark.

More info in read.me

Integration testing combines individual 0chain modules and test them as a group. Integration testing evaluates the compliance of a system for specific functional requirements and usually occurs after unit testing .

For integration testing, A conductor which is RPC(Remote Procedure Call) server is implemented to control behaviour of nodes .To know more about the conductor refer to the conductor documentation

A conductor requires the nodes to be built in a certain order to control them during the tests. A config file is defined in conductor.config.yaml which contains important details such as details of all nodes used and custom commands used in integration testing.

For running multiple test cases,conductor uses a test suite which contains multiple sets of tests .A test suites can be categorized into 3 types of tests

standard tests - Checks whether chain continue to function properly despite bad miner and sharder participants

view-change tests - Checks whether addition and removal of nodes is working

.blobber tests - Checks whether storage functions continue to work properly despite bad or lost blobber, and confirms expected storage function failures

Below is an example of conductor test suite.

# Under `enable` is the list of sets that will be run.
enable:
  - "Miner down/up"
  - "Blobber tests"

# Test sets defines the test cases it covers.
sets:
  - name: "Miner down/up"
    tests:
      - "Miner: 50 (switch to contribute)"
      - "Miner: 100 (switch to share)"
  - name: "Blobber tests"
    tests:
      - "All blobber tests"

# Test cases defines the execution flow for the tests.
tests:
  - name: "Miner: 50 (switch to contribute)"
    flow:
    # Flow is a series of directives.
    # The directive can either be built-in in the conductor
    # or custom command defined in "conductor.config.yaml"
      - set_monitor: "sharder-1" # Most directive refer to node by name, these are defined in `conductor.config.yaml`
      - cleanup_bc: {} # A sample built-in command that triggers stop on all nodes and clean up.
      - start: ['sharder-1']
      - start: ['miner-1', 'miner-2', 'miner-3']
      - wait_phase:
          phase: 'contribute'
      - stop: ['miner-1']
      - start: ['miner-1']
      - wait_view_change:
          timeout: '5m'
          expect_magic_block:
            miners: ['miner-1', 'miner-2', 'miner-3']
            sharders: ['sharder-1']
  - name: "Miner: 100 (switch to share)"
    flow:
    ...
  - name: "All blobber tests"
    flow:
      - command:
          name: 'build_test_blobbers' # Sample custom command that executes `build_test_blobbers`
    ...
...

Docker and Git must be installed to run the tests .

Install Git using the following command:

sudo apt install git

Docker installation instructions can be found here.

Clone the 0chain repository:

git clone https://github.com/0chain/0chain.git

Build miner docker image for integration test

(cd 0chain && ./docker.local/bin/build.miners-integration-tests.sh)

Build sharder docker image for integration test

(cd 0chain && ./docker.local/bin/build.sharders-integration-tests.sh)

NOTE: The miner and sharder images are designed for integration tests only. If wanted to run chain normally, rebuild the original images.

(cd 0chain && ./docker.local/bin/build.sharders.sh && ./docker.local/bin/build.miners.sh)

Confirm that view change rounds are set to 50 on 0chain/docker.local/config.yaml

    start_rounds: 50
    contribute_rounds: 50
    share_rounds: 50
    publish_rounds: 50
    wait_rounds: 50

Run miners test

(cd 0chain && ./docker.local/bin/start.conductor.sh miners)

Run sharders test

(cd 0chain && ./docker.local/bin/start.conductor.sh sharders)

1. These 2 scripts should be run with view_change: false in 0chain/docker.local/config.yaml 1.1. (cd 0chain && ./docker.local/bin/start.conductor.sh no-view-change.byzantine) 1.2. (cd 0chain && ./docker.local/bin/start.conductor.sh no-view-change.fault-tolerance)
2. Set view_change: true in 0chain/docker.local/config.yaml for the following 2 scripts 2.1. (cd 0chain && ./docker.local/bin/start.conductor.sh view-change.byzantine) 2.2. (cd 0chain && ./docker.local/bin/start.conductor.sh view-change.fault-tolerance*)

Refer to conductor documentation

New tests can be easily added to the conductor check Updating conductor tests in the conductor documentation for more information.

Check Temporarily disabling tests in the conductor documentation for more information

Check the supported directives in the conductor documentation for more information.

Check Custom Commands in the conductor documentation for more information

To generate swagger documentation you need go-swagger installed, visit https://goswagger.io/install.html for details.

You then need to run the makefile

make swagger

The documentation will be in docs/swagger.md and docs/swagger.yaml.