Fabric基于docker方式、多机kafka配置部署详细过程。
更多区块链技术与应用分类：
基于docker方式的多机kafka的Fabric网络搭建步骤如下。
分布式节点规划：
node1: kafka1 zookeeper1 orderer1
node2: kafka2 zookeeper2 orderer2 peer1
node3: kafka3 zookeeper3 orderer3 peer2
node4: kafka4 peer3
每台节点网络ip配置(vim /etc/hosts)
172.27.34.201 orderer1.trace.com zookeeper1 kafka1
172.27.34.202 orderer2.trace.com zookeeper2 kafka2 peer0.org1.trace.com org1.trace.com
172.27.34.203 orderer3.trace.com zookeeper3 kafka3 peer1.org2.trace.com org2.trace.com
172.27.34.204 kafka4 peer2.org3.trace.com org3.trace.com
环境需求
fabric v1.1
docker(17.06.2-ce或更高)
扫描二维码关注公众号，回复：
12013241 查看本文章
docker-compose(1.14.0或更高)
go( 1.9.x或更高)
docker安装
删除之前安装的
sudo yum remove docker docker-common docker-selinux docker-engine
安装一些依赖
sudo yum install -y yum-utils device-mapper-persistent-data lvm2
下载安装rpm包
sudo yum install https://download.docker.com/linux/centos/7/x86_64/stable/Packages/docker-ce-17.06.2.ce-1.el7.centos.x86_64.rpm
go更新：
rm -rf /usr/local/go/
wget https://dl.google.com/go/go1.12.linux-amd64.tar.gz
docker镜像下载
docker pull registry.docker-cn.com/hyperledger/fabric-peer:x86_64-1.1.0
docker pull registry.docker-cn.com/hyperledger/fabric-orderer:x86_64-1.1.0
docker pull registry.docker-cn.com/hyperledger/fabric-tools:x86_64-1.1.0
docker pull hyperledger/fabric-couchdb
docker pull hyperledger/fabric-ca
docker pull hyperledger/fabric-ccenv
docker pull hyperledger/fabric-baseos
docker pull hyperledger/fabric-kafka
docker pull hyperledger/fabric-zookeeper
改标签：
docker tag b7bfddf508bc hyperledger/fabric-tools:latest
docker tag ce0c810df36a hyperledger/fabric-orderer:latest
docker tag b023f9be0771 hyperledger/fabric-peer:latest
删除相同的：
docker rmi registry.docker-cn.com/hyperledger/fabric-tools:x86_64-1.1.0
docker rmi registry.docker-cn.com/hyperledger/fabric-orderer:x86_64-1.1.0
docker rmi registry.docker-cn.com/hyperledger/fabric-peer:x86_64-1.1.0
保存镜像并发送到其他节点：
=>node1:
mkdir -p /data/fabric-images-1.1.0-release
docker save 5b31d55f5f3a > /data/fabric-images-1.1.0-release/fabric-ccenv.tar
docker save 1a804ab74f58 > /data/fabric-images-1.1.0-release/fabric-ca.tar
docker save d36da0db87a4 > /data/fabric-images-1.1.0-release/fabric-zookeeper.tar
docker save a3b095201c66 > /data/fabric-images-1.1.0-release/fabric-kafka.tar
docker save f14f97292b4c > /data/fabric-images-1.1.0-release/fabric-couchdb.tar
docker save 75f5fb1a0e0c > /data/fabric-images-1.1.0-release/fabric-baseos.tar
docker save b7bfddf508bc > /data/fabric-images-1.1.0-release/fabric-tools.tar
docker save ce0c810df36a > /data/fabric-images-1.1.0-release/fabric-orderer.tar
docker save b023f9be0771 > /data/fabric-images-1.1.0-release/fabric-peer.tar
docker save b023f9be0771 > /data/fabric-images-1.1.0-release/fabric-peer.tar
docker save b023f9be0771 > /data/fabric-images-1.1.0-release/fabric-peer.tar
cd /data/fabric-images-1.1.0-release
scp -r ./* [email protected]:/tmp/docker/fabric-images/
在node2，node3，node4：
docker load < fabric-baseos.tar
docker tag 75f5fb1a0e0c hyperledger/fabric-baseos:latest
docker load < fabric-ca.tar
docker tag 1a804ab74f58 hyperledger/fabric-ca:latest
docker load < fabric-ccenv.tar
node1：
操作在/traceabilityProject下
下载二进制文件(v1.1版本)：
一.配置文件设置：
生成节点所需配置文件：
./bin/cryptogen generate --config=./crypto-config_kafka.yaml
二.生成创始区块：
变更部分(集群地址)：
将该文件放入/traceabilityProject内，然后运行在该目录下：
(必须改名为configtx.yaml文件)
export FABRIC_CFG_PATH=$PWD
mkdir ./channel-artifacts
./bin/configtxgen -profile TestOrgsOrdererGenesis -outputBlock ./channel-artifacts/orderer.genesis.block
三.生成通道配置文件：
./bin/configtxgen -profile TestOrgsChannel -outputCreateChannelTx ./channel-artifacts/tracechannel.tx -channelID tracechannel
四.启动环境与文件配置
说明：由于每次容器销毁后，网络和数据就都不存在了，这对于生产环境来说是无法接受的，因此，所以考虑数据持久化。
需要持久化数据的组件：Orderer、Peer、Kafka、zookeeper
在启动文件中加入：
Orderer：
environment: - ORDERER_FILELEDGER_LOCATION=/traceabilityProject/orderer/fileLedger volumes: - /traceabilityProject/orderer/fileLedger:/traceabilityProject/orderer/fileLedger
Peer：
environment: - CORE_PEER_FILESYSTEMPATH=/traceabilityProject/peer/production volumes: - /traceabilityProject/peer/production:/traceabilityProject/peer/production
Kafka：
environment: #Kafka数据也要持久化 - KAFKA_LOG.DIRS=/traceabilityProject/kafka/kafka-logs #数据挂载路径 volumes: - /traceabilityProject/kafka/kafka-logs:/traceabilityProject/kafka/kafka-logs
zookeeper：
volumes: #因为zookeeper默认数据存储路径为/data及/datalog - /traceabilityProject/zookeeper1/data:/data - /traceabilityProject/zookeeper1/datalog:/datalog
详细配置如下面yaml文档所示
node1:
再创建kafka、orderer、zookeeper三个文件夹来持久存储数据
docker-compose -f ./node1-docker-compose-up.yaml up -d
停止容器则使用 ：
docker-compose -f ./node1-docker-compose-up.yaml down
注：
在配置文件中volume中目录中，复制生成的相应配置文件(这些目录不能为空！)
(
创世区块目录要挂载文件而非文件夹，但docker-compose每次生成的都是文件夹，手动更改为创世区块文件：
- ../channel-artifacts/orderer.genesis.block:/traceabilityProject/orderer.genesis.block
)
将工作目录/traceabilityProject复制到其他节点：
scp -r /traceabilityProject/ [email protected]:/
scp -r /traceabilityProject/ [email protected]:/
scp -r /traceabilityProject/ [email protected]:/
node2：
创建kafka、orderer、zookeeper、peer四个文件夹，来持久存储数据
docker-compose -f ./node2-docker-compose-up.yaml up -d
docker-compose -f ./node2-docker-compose-up.yaml down --remove-orphans
node2-docker-compose-up.yaml
node3：
创建kafka、orderer、peer三个文件夹，来持久存储数据
node4：
创建kafka、peer两个文件夹，来持久存储数据
注意：
只有所有的zookeeper和kafka都启动后(四份文件同时启动,会有一些延时等待时间)，orderer节点才不会宕掉。
(否则报错：
Cannot post CONNECT message = circuit breaker is open)
五.Peer节点操作(交易、通道、链码)
peer1(node2)：
(1)进入容器
docker exec -it cli bash
查看容器工作目录下是否有文件，若没有则手动从主机拷贝到docker 容器中：
(2)创建通道
容器中
cd /traceabilityProject/peer/channel-artifacts
peer channel create -o orderer1.trace.com:7050 -c tracechannel -t 50 -f ./tracechannel.tx
(3)发送tracechannel.block给其他节点：
由于/traceabilityProject/peer/channel-artifacts是容器的挂载目录，所以物理机中已有该文件。
向两个peer节点拷贝：
退出容器，进入主机目录：/traceabilityProject/channel-artifacts
scp -r tracechannel.block [email protected]:/traceabilityProject/channel-artifacts
scp -r tracechannel.block [email protected]:/traceabilityProject/channel-artifacts
由于/traceabilityProject/channel-artifacts是cli容器中的volumes挂载点，在容器中。
(4)加入通道
再次进入容器目录/traceabilityProject/peer/channel-artifacts：
peer channel join -b tracechannel.block
(
完全重启指令：
docker-compose -f node1-docker-compose-up.yaml down --volumes --remove-orphans docker volume prune
)
(即使没有设置锚节点的情况下，整个Fabric网络仍然是能正常运行的)
(5)安装链码
peer chaincode install -n cc_producer -v 1.0 -p github.com/hyperledger/fabric/examples/chaincode/go/tireTraceability-Demo/main/
(6)初始化
peer chaincode instantiate -o orderer1.trace.com:7050 -C tracechannel -n cc_producer -v 1.0 -c '{"Args":["init"]}' -P "OR ('Org1MSP.member','Org2MSP.member','Org3MSP.member')"
pee2(node3):
docker exec -it cli bash
cd channel-artifacts/
peer channel join -b tracechannel.block
peer chaincode install -n cc_agency -v 1.0 -p github.com/hyperledger/fabric/chaincode/go/traceability/src/origin_agency/main/
peer chaincode instantiate -o orderer1.trace.com:7050 -C tracechannel -n cc_agency -v 1.0 -c '{"Args":["init"]}' -P "OR ('Org1MSP.member','Org2MSP.member','Org3MSP.member')"
peer3(node4):
docker exec -it cli bash
cd channel-artifacts/
peer channel join -b tracechannel.block
peer chaincode install -n cc_producer -v 1.0 -p github.com/hyperledger/fabric/chaincode/go/traceability/src/origin_producer/main/
peer chaincode install -n cc_agency -v 1.0 -p github.com/hyperledger/fabric/chaincode/go/traceability/src/origin_agency/main/
peer chaincode install -n cc_retailer -v 1.0 -p github.com/hyperledger/fabric/chaincode/go/traceability/src/origin_retailer/main/
peer chaincode instantiate -o orderer1.trace.com:7050 -C tracechannel -n cc_retailer -v 1.0 -c '{"Args":["init","A","B","C"]}' -P "OR ('Org1MSP.member','Org2MSP.member','Org3MSP.member')"
(每次重启容器，需再走一遍流程，加入通道、安装、，retailer上安装三个链码，但只初始化cc_retailer)
完全重启指令：
docker-compose -f node1-docker-compose-up.yaml down --volumes --remove-orphans
docker-compose -f node2-docker-compose-up.yaml down --volumes --remove-orphans
docker-compose -f node3-docker-compose-up.yaml down --volumes --remove-orphans
docker-compose -f node4-docker-compose-up.yaml down --volumes --remove-orphans docker volume prune
并删除持久化的数据
rm -rf kafka/* orderer/* zookeeper1/*
rm -rf kafka/* orderer/* peer/* zookeeper2/*
rm -rf kafka/* orderer/* peer/* zookeeper3/*
rm -rf kafka/* peer/*
关机前关闭容器时：
docker-compose -f /traceabilityProject/startup/node1-docker-compose-up.yaml down
docker-compose -f /traceabilityProject/startup/node2-docker-compose-up.yaml down
docker-compose -f /traceabilityProject/startup/node3-docker-compose-up.yaml down
docker-compose -f /traceabilityProject/startup/node4-docker-compose-up.yaml down