Docker Community Forums

Share and learn in the Docker community.

UCP : Can not add new manager in swarm cluster

dockercloud

(Questel) #1

Hello,

Actually I am testing trial swarm/ucp version based on our local centos7 cluster with firewall off. According to docker doc, I have uninstall docker ce to install docker ee.

docker --version -> Docker version 17.06.2-ee-6, build e75fdb8

I have successfully “init” first docker swarm manager. I have uploaded trial licence and can browse web UI. Also I can successfully add worker and can see it in web UI with green light.

However after many tries, I can not add new manager. Web UI reports “One or more required ports is unavailable” for this node". After copy and paste, docker logs reports error :

docker swarm join --listen-addr xxx.xxx..xxx.xxx:2377 --token SWMTKN-1-00a7q0qqway027wdcl2p13loooeyhs9ei0qtjf7jdp1p29wefc-ceqxfcz851e6edfhayje0v8ls xxx.xxx.xxx.xxx:2377
This node joined a swarm as a manager.

docker logs dc9bb35f7f78
{"level":"info","msg":"Initializing UCP Agent ...","time":"2018-01-23T13:45:50Z"}
{"level":"info","msg":"Connecting to local Docker daemon.","time":"2018-01-23T13:45:50Z"}
{"level":"info","msg":"Loading UCP Configuration file.","time":"2018-01-23T13:45:50Z"}
{"level":"info","msg":"UCP agent initialization complete.","time":"2018-01-23T13:45:50Z"}
{"level":"info","msg":"Beginning state difference detection loop...","time":"2018-01-23T13:45:50Z"}

Then

{"level":"info","msg":"detected state difference for component: Docker Proxy","time":"2018-01-23T13:46:53Z"}
{"level":"info","msg":"Detected difference in desired and current state of UCP on this node.","time":"2018-01-23T13:46:53Z"}
{"level":"error","msg":"port check for Swarm-Classic Manager component unsuccessful: Error while checking for available ports: The following required ports are already in use on your host - 2376.  You may specify an alternative port number to 2376 with the --swarm-port argument.","time":"2018-01-23T13:47:00Z"}
{"level":"error","msg":"port check for Concurrent [Swarm-Classic Manager eNZi API eNZi Worker prometheus s390xservice windowsservice] component unsuccessful: Error while checking for available ports: The following required ports are already in use on your host - 2376.  You may specify an alternative port number to 2376 with the --swarm-port argument.","time":"2018-01-23T13:47:00Z"}
{"level":"error","msg":"Port check failed: Error while checking for available ports: The following required ports are already in use on your host - 2376.  You may specify an alternative port number to 2376 with the --swarm-port argument.","time":"2018-01-23T13:47:00Z"}
Error while checking for available ports: The following required ports are already in use on your host - 2376.  You may specify an alternative port number to 2376 with the --swarm-port argument.

Boring stuff to troubleshoot/reproduce error is I have to re init from scratch because swarm cluster is lost.

Thank you in advance for your help
Nicolas