Chaincode should be re-deployed after restarting peer node

[yingfeng]

Description

Chaincode runs as a seperate docker container. After it has been successfully deployed, it could not connect to peer anymore after peer had been restarted, unless the chaincode is re-deployed.

Describe How to Reproduce

• Start the peer node using docker on baremetal

docker run --rm -it --name peer1 -e CORE_VM_ENDPOINT=http://172.17.0.1:25 -e CORE_PEER_ID=vp0 -e CORE_PEER_ADDRESSAUTODETECT=true hyperledger-peer peer node start

• Deploy a chaincode

{
  "jsonrpc": "2.0",
  "method": "deploy",
  "params": {
    "type": 1,
    "chaincodeID":{
        "path":"github.com/hyperledger/fabric/examples/chaincode/go/chaincode_example02"
    },
    "ctorMsg": {
        "function":"init",
        "args":["a", "1000", "b", "2000"]
    }
  },
  "id": 1
}

• Restart the peer node
• Query the chaincode

{
  "jsonrpc": "2.0",
  "method": "query",
  "params": {
      "type": 1,
      "chaincodeID":{
          "name":"7b07c59e9b9405c1aef33493b63b9a766d9bb836989ded1730052de650aa8ce5654274d148ceff96a4e5bd43bca26aba099f55c400e4befdc8b2ee4c0a94e30b",
          "path":"github.com/hyperledger/fabric/examples/chaincode/go/chaincode_example02"
      },
      "ctorMsg": {
         "function":"query",
         "args":["a"]
      }
  },
  "id": 5
}

When using docker inspect to find the status differences, I've found that {{ .NetworkSettings.Networks.host.EndpointID }} has been set, while after restarting peer node, that field is empty. It's reasonable to enhance the connection management between chaincode and peer, such that the chaincode is not needed to re-deploy whenever peer had been restarted.

[corecode]

@muralisrini