Getting started with Hyperledger on Kubernetes

Why?

Recently, I have been following the Hyperledger project, and Fabric in particular, with fair bit of interest. The current deployment process1 for Fabric Starter Kit uses Docker Swarm.

Kubernetes is a leading platform for automating deployment, scaling, and operations of containerised applications.

Using Kubernetes instead of Docker Swarm would allow Hyperledger Fabric to leverage features like:

  1. Automatic binpacking
  2. Horizontal scaling
  3. Automated rollouts and rollbacks
  4. Storage orchestration
  5. Self-healing
  6. Service discovery and load balancing
  7. Secret and configuration management
  8. Batch execution

So this article looks at a simple way to deploy Fabric Starter Kit on a Kubernetes platform. This would be the first step in getting a production-grade Fabric deployment on Kubernetes.

With developer as the primary audience, we will use OpenShift as the underlying platform; as Kubernetes comes bundled in with OpenShift.

What?

Fabric application is composed of functional components distributed as linux containers using the Docker container packaging format.

Kubernetes is an open-source platform for automating deployment, scaling, and operations of application containers across clusters of hosts, providing container-centric infrastructure.

OpenShift Container Platform – OCP is based on top of Linux containers and the Kubernetes container cluster manager. OCP adds developer and operational management centric tools to enable rapid application development, easy deployment and scaling, and long-term lifecycle maintenance for small and large teams and applications.

How?

Environment

I am using <kbd>macOS Sierra</kbd> for development, with following configuration:

Steps to deploy fabric-starter-kit on OpenShift/Kubernetes

Within a <kbd>terminal</kbd> perform following steps.

  1. Bring up OpenShift service
oc cluster up

First time you execute this command it will take sometime as the service might download required OpenShift images
Note the information within –Server Information section

-- Server Information ...
OpenShift server started.
The server is accessible via web console at:
https://192.168.0.7:8443
You are logged in as:
User: developer
Password: developer
To login as administrator:
oc login -u system:admin
  1. Allow the application to run as <kbd>root</kbd> or within a privileged container.
    > Allowing application containers to run as <kbd>root</kbd> by default should be strongly discouraged2
  • Login as an administrator
oc login -u system:admin
oc adm policy add-scc-to-user privileged -z default -n myproject
  • Logout as administrator and log back in as the default user
oc login -u developer
  1. Download the docker-compose file for the starter-kit
wget -O docker-compose-fabric-starter-kit.yml https://raw.githubusercontent.com/hyperledger/fabric/master/examples/sdk/node/docker-compose.yml
  1. Update the downloaded YAML file to ensure appropriate ports are exposed4.
# updated snippet of docker-compose-fabric-starter-kit.yml
membersrvc:
ports:
- &quot;7054:7054&quot;

peer:
ports:
- &quot;7051:7051&quot;
  1. Import the docker-compose application into OpenShift
oc import docker-compose -f docker-compose-fabric-starter-kit.yml
  1. Allow member service to be executed with root status
oc patch dc membersrvc -p &#039;{&quot;spec&quot;:{&quot;template&quot;:{&quot;spec&quot;:{&quot;containers&quot;:[{&quot;name&quot;:&quot;membersrvc&quot;,&quot;securityContext&quot;:{&quot;privileged&quot;:true}}]}}}}&#039;
  1. Allow validating peer service vp0 to be executed with root status
oc patch dc vp0 -p &#039;{&quot;spec&quot;:{&quot;template&quot;:{&quot;spec&quot;:{&quot;containers&quot;:[{&quot;name&quot;:&quot;vp0&quot;,&quot;securityContext&quot;:{&quot;privileged&quot;:true}}]}}}}&#039;
  1. Give the system couple of minutes to startup the required containers. Then ensure that all the services are up and running.
oc get pods

&gt;Result should be similar to below:

NAME READY STATUS RESTARTS AGE
membersrvc-2-3ft1y 1/1 Running 0 16m
peer-2-3p43z 1/1 Running 0 15m
starter-1-h8z1d 1/1 Running 1 16m
  1. <kbd>membersrvc-2-3ft1y</kbd> is the member service
  2. <kbd>peer-2-3p43z</kbd> is the validating peer service
  3. <kbd>starter-1-h8z1d</kbd> is the fabric-starter-kit service

  4. Deploy and execute the sample chaincode5 using Node.js Client SDK within the fabric-starter-kit

oc exec starter-1-h8z1d node app

&gt;Result should be similar to below:

*** starting HFC sample ***
member services address =membersrvc:7054
peer address =peer:7051
DEPLOY_MODE=dev
enrolling user admin ...
Enrolled JohnDoe successfully
deploying chaincode; please wait ...
deploy complete; results: {&quot;uuid&quot;:&quot;mycc&quot;,&quot;chaincodeID&quot;:&quot;mycc&quot;}
invoke chaincode ...
invoke submitted successfully; results={&quot;uuid&quot;:&quot;85782c59-e192-499f-8a5d-108c4423133d&quot;}
invoke completed successfully; results={&quot;result&quot;:&quot;Tx 85782c59-e192-499f-8a5d-108c4423133d complete&quot;}
querying chaincode ...
query completed successfully; results={&quot;result&quot;:{&quot;type&quot;:&quot;Buffer&quot;,&quot;data&quot;:[57,55]}}
  1. Investigate logs for the peer validating service
oc logs peer-2-3p43z

&gt;Result should be similar to:

04:06:46.438 [peer] ProcessTransaction -&gt; DEBU 1128 ProcessTransaction processing transaction txid = mycc
04:06:46.438 [peer] ProcessTransaction -&gt; DEBU 1129 Verifying transaction signature mycc
04:06:46.438 [crypto] Debugf -&gt; DEBU 112a [validator.vp] Tx confdential level [PUBLIC].
04:06:46.439 [peer] sendTransactionsToLocalEngine -&gt; DEBU 112b Marshalling transaction CHAINCODE_DEPLOY to send to local engine
04:06:46.439 [peer] sendTransactionsToLocalEngine -&gt; DEBU 112c Sending message CHAIN_TRANSACTION with timestamp seconds:1477368406 nanos:439541357 to local engine
04:06:46.439 [consensus/noops] RecvMsg -&gt; DEBU 112d Handling Message of type: CHAIN_TRANSACTION
04:06:46.439 [consensus/noops] broadcastConsensusMsg -&gt; DEBU 112e Broadcasting CONSENSUS
04:06:46.439 [peer] Broadcast -&gt; DEBU 112f Broadcast took 3.096µs
04:06:46.439 [consensus/noops] RecvMsg -&gt; DEBU 1130 Sending to channel tx uuid: mycc
04:06:46.559 [peer] ensureConnected -&gt; DEBU 1131 Touch service indicates no dropped connections
04:06:46.559 [peer] ensureConnected -&gt; DEBU 1132 Connected to: []
04:06:46.559 [peer] ensureConnected -&gt; DEBU 1133 Discovery knows about: []
04:06:47.440 [consensus/noops] handleChannels -&gt; DEBU 1134 Process block due to time
04:06:47.440 [consensus/noops] processTransactions -&gt; DEBU 1135 Starting TX batch with timestamp: seconds:1477368407 nanos:440946960
04:06:47.441 [consensus/noops] processTransactions -&gt; DEBU 1136 Executing batch of 1 transactions with timestamp seconds:1477368407 nanos:440946960
04:06:47.441 [crypto] Debugf -&gt; DEBU 1137 [validator.vp] Tx confdential level [PUBLIC].
04:06:47.441 [chaincode] Deploy -&gt; DEBU 1138 user runs chaincode, not deploying chaincode
04:06:47.442 [state] TxBegin -&gt; DEBU 1139 txBegin() for txId [mycc]
04:06:47.442 [chaincode] Launch -&gt; DEBU 113a chaincode is running(no need to launch) : mycc
04:06:47.444 [state] TxFinish -&gt; DEBU 113b txFinish() for txId [mycc], txSuccessful=[true]
04:06:47.444 [state] GetHash -&gt; DEBU 113c Enter - GetHash()
04:06:47.445 [buckettree] ComputeCryptoHash -&gt; DEBU 113d Enter - ComputeCryptoHash()
04:06:47.445 [buckettree] ComputeCryptoHash -&gt; DEBU 113e Returing existing crypto-hash as recomputation not required
04:06:47.446 [state] GetHash -&gt; DEBU 113f Exit - GetHash()
04:06:47.447 [consensus/noops] processTransactions -&gt; DEBU 1140 Committing TX batch with timestamp: seconds:1477368407 nanos:440946960
04:06:47.447 [state] GetHash -&gt; DEBU 1141 Enter - GetHash()
04:06:47.448 [buckettree] ComputeCryptoHash -&gt; DEBU 1142 Enter - ComputeCryptoHash()
04:06:47.448 [buckettree] ComputeCryptoHash -&gt; DEBU 1143 Returing existing crypto-hash as recomputation not required
04:06:47.448 [state] GetHash -&gt; DEBU 1144 Exit - GetHash()
04:06:47.449 [indexes] addIndexDataForPersistence -&gt; DEBU 1145 Indexing block number [5] by hash = [6132347bf82b609e6f28c5c8c98627f36cc4b7b64c713bc36af79aafc9ef4d8851ec66c28d7eb6bfec5e04d811e39cd6377f2cda8b2dce55bd6628b41e5fbd61]
04:06:47.449 [state] AddChangesForPersistence -&gt; DEBU 1146 state.addChangesForPersistence()...start
04:06:47.449 [state] AddChangesForPersistence -&gt; DEBU 1147 Adding state-delta corresponding to block number[5]
04:06:47.449 [state] AddChangesForPersistence -&gt; DEBU 1148 Not deleting previous state-delta. Block number [5] is smaller than historyStateDeltaSize [500]
04:06:47.450 [state] AddChangesForPersistence -&gt; DEBU 1149 state.addChangesForPersistence()...finished
04:06:47.450 [ledger] resetForNextTxGroup -&gt; DEBU 114a resetting ledger state for next transaction batch
04:06:47.450 [buckettree] ClearWorkingSet -&gt; DEBU 114b Enter - ClearWorkingSet()
04:06:47.451 [ledger] CommitTxBatch -&gt; DEBU 114c There were some erroneous transactions. We need to send a &#039;TX rejected&#039; message here.
04:06:47.451 [consensus/handler] CommitTxBatch -&gt; DEBU 114d Committed block with 1 transactions, intended to include 1
04:06:47.451 [consensus/noops] getBlockData -&gt; DEBU 114e Preparing to broadcast with block number 6
04:06:47.451 [consensus/noops] getBlockData -&gt; DEBU 114f Got the delta state of block number 6
04:06:47.451 [consensus/noops] notifyBlockAdded -&gt; DEBU 1150 Broadcasting Message_SYNC_BLOCK_ADDED to non-validators
04:06:47.452 [peer] Broadcast -&gt; DEBU 1151 Broadcast took 3.795µs

This proves how easy it is to start developing chaincode applications4 to run on Fabric network, particularly when using OpenShift as a deployment target — with Kubernetes built in.

Next steps to consider

  • Use the fabric-starter-kit and run it as a standalone app in an external OpenShift project
  • Get the peer and member services to run with non-root privilege

  1. The current starter-kit uses an application defined within docker-compose.yml 
  2. Architecting containers: First article in a series of four(4) 
  3. <kbd>myproject</kbd> is the project created by default. <kbd>oc login -u developer</kbd> command tells you which project you are using. 
  4. An example of the updated YAML file is (available on github)[https://raw.githubusercontent.com/finiteloopme/hyperledger-on-kubernetes/master/docker-compose-fabric-starter-kit.yml]. 
  5. Chaincode is an application-level code stored on the ledger as part of a transaction. 

Join the Red Hat Developer Program (it’s free) and get access to related cheat sheets, books, and product downloads.


For more information about Red Hat OpenShift and other related topics, visit: OpenShift, OpenShift Online.

Share