Port error when setting up Dev mode of Hyperledger Fabric - docker

Question

I'm setting up the development environment following the instructions on Hyperledger fabric's official website:
https://hyperledger-fabric.readthedocs.io/en/latest/peer-chaincode-devmode.html
I have started the orderer successfully using:
ORDERER_GENERAL_GENESISPROFILE=SampleDevModeSolo orderer
This command didn't work at first but it worked after I cd fabric/sampleconfig
2020-12-21 11:23:15.084 CST [orderer.common.server] Main -> INFO 009 Starting orderer: Version: 2.3.0 Commit SHA: dc2e59b3c Go version: go1.15.6 OS/Arch: darwin/amd64
2020-12-21 11:23:15.084 CST [orderer.common.server] Main -> INFO 00a Beginning to serve requests
but when I start the peer using:
export PATH=$(pwd)/build/bin:$PATH
export FABRIC_CFG_PATH=$(pwd)/sampleconfig
export FABRIC_LOGGING_SPEC=chaincode=debug
export CORE_PEER_CHAINCODELISTENADDRESS=0.0.0.0:7052
peer node start --peer-chaincodedev=true
An error is spotted:
FABRIC_LOGGING_SPEC=chaincode=debug
CORE_PEER_CHAINCODELISTENADDRESS=0.0.0.0:7052
peer node start --peer-chaincodedev=true
2020-12-21 11:25:13.047 CST [nodeCmd] serve -> INFO 001 Starting peer: Version: 2.3.0 Commit SHA: dc2e59b3c Go version: go1.15.6 OS/Arch: darwin/amd64 Chaincode: Base Docker Label: org.hyperledger.fabric Docker Namespace: hyperledger
2020-12-21 11:25:13.048 CST [peer] getLocalAddress -> INFO 002 Auto-detected peer address: 10.200.83.208:7051
2020-12-21 11:25:13.048 CST [peer] getLocalAddress -> INFO 003 Host is 0.0.0.0 , falling back to auto-detected address: 10.200.83.208:7051 Error: failed to initialize operations subsystem: listen tcp 127.0.0.1:9443: bind: address already in use
this is the error:
Error: failed to initialize operations subsystem: listen tcp 127.0.0.1:9443: bind: address already in use
I checked this issue and it seems this happens because the peer node is using the same port 9443 as the orderer node for the same service. How can I get the two nodes running separately? It seems the docker is running as well.

Answer 1

If you see your error, you can easily follow
Error: failed to initialize operations subsystem: listen tcp 127.0.0.1:9443: bind: address already in use
It is said that the 9443 port is already in use.
It seems that you are not running the orderer and peer as separate containers on the docker-based virtual network, but running on the host pc.
This eventually seems to conflict with two servers requesting one port 9443 on your pc.\
Referring to the configuration below of fabric-2.3/sampleconfig, you can see that each port 9443 is assigned to the server. Assigning one of them to the other port solves this.
fabric-2.3/sampleconfig/orderer.yaml
configuration of orderer
# orderer.yaml
...
Admin:
# host and port for the admin server
ListenAddress: 127.0.0.1:9443
...
fabric-2.3/sampleconfig/core.yaml
configuration of peer
# core.yaml
...
operations:
# host and port for the operations server
# listenAddress: 127.0.0.1:9443
listenAddress: 127.0.0.1:10443
...

Answer 2

This is not a direct answer to the port mapping / collision issue, but we've had great success using the new Kubernetes Test Network as a development platform running on a local system with a virtual Kubernetes cluster running in KIND (Kubernetes in Docker).
In this mode, applications can be developed using the Gateway client (exposed via a port forward or ingress), and smart contracts running As a Service can be launched either in the cluster OR run on the local host OS in a container, binary, or launched in a debugger.
The documentation for the development setup is still sparse, but we'd love to hear feedback on the overall approach, as it offers an exponentially better experience for working with a test network in a development context. In general the process of "port juggling" with Compose is no longer relevant when working on a local Kubernetes cluster. In this mode, you can run services on the host network, instructing peers/orderers/etc. to connect to the remote process running on the host OS.