Docker.github.io: Failed to join swarm as documented

[dbackeus]

File: get-started/part4.md, CC @johndmulhausen

After adding the --advertise-addr argument to init the swarm on myvm1 I got the expected output:

$ docker-machine ssh myvm1 "docker swarm init --advertise-addr 192.168.99.100:2376"
Swarm initialized: current node (s8b86jmd6gi8dk9ftx07cyeop) is now a manager.

To add a worker to this swarm, run the following command:

    docker swarm join \
    --token SWMTKN-1-0mvkhaacjfoppv7x8t6ffiue45zs2plhlsngljk08p078qis5j-eg3news5k68kgmhby38wbsyqu \
    192.168.99.100:2376

However when trying to join it from myvm2 I get the following error:

$ docker-machine ssh myvm2 "docker swarm join --token SWMTKN-1-0mvkhaacjfoppv7x8t6ffiue45zs2plhlsngljk08p078qis5j-eg3news5k68kgmhby38wbsyqu 192.168.99.100:2376"
Error response from daemon: rpc error: code = 13 desc = connection error: desc = "transport: remote error: tls: bad certificate"
exit status 1

I'm running macOS Sierra 10.12.4 on a late 2013 15" Macbook Pro with Virtualbox 5.1.20.

Docker version is:

$ docker version
Client:
 Version:      17.03.1-ce
 API version:  1.27
 Go version:   go1.7.5
 Git commit:   c6d412e
 Built:        Tue Mar 28 00:40:02 2017
 OS/Arch:      darwin/amd64

Server:
 Version:      17.03.1-ce
 API version:  1.27 (minimum version 1.12)
 Go version:   go1.7.5
 Git commit:   c6d412e
 Built:        Fri Mar 24 00:00:50 2017
 OS/Arch:      linux/amd64
 Experimental: true

[johndmulhausen]

The doc says to make sure you're using port 2377: https://docs.docker.com/get-started/part4/#create-a-cluster

If you run these commands again using that port, does it work? It's not that you couldn't make it work with 2376, but I think a config change would need to happen. Let me know.

[siliconmeadow]

@johndmulhausen I'm cross posting my comment from here: https://github.com/docker/machine/issues/4064

It's confusing though. The tutorial is correct - port 2377 is the port that works, but while running the commands to initialise the docker swarm a user sees:

$ docker-machine ssh myvm1 "docker swarm init --advertise-addr 192.168.99.100:2376"
Swarm initialized: current node (xpjllelqsli4xk2m9f8plzzc7) is now a manager.

To add a worker to this swarm, run the following command:

    docker swarm join \
    --token SWMTKN-1-2xw6b7rsyn5ou92rejukqvtayar3t9jz68eqd7y0m367psu7vf-dcdk4bd7lhashxx9ki460su7h \
    192.168.99.100:2376

To add a manager to this swarm, run 'docker swarm join-token manager' and follow the instructions.

Note: Port 2376

[siliconmeadow]

I've never worked with Go, so I could be wrong - I'm looking though the code in the cli repo and found this:

https://github.com/siliconmeadow/cli/blob/master/cli/command/swarm/join_token.go#L111

[dbackeus]

Just a heads up that I wiped my docker setup and will likely not get back to actually trying this out so can't confirm the port config theory.

[vorpus]

Same issue, used port 2377 as instructed in the tutorial.

$ docker version
Client:
 Version:      17.03.1-ce
 API version:  1.27
 Go version:   go1.7.5
 Git commit:   c6d412e
 Built:        Tue Mar 28 00:40:02 2017
 OS/Arch:      darwin/amd64

Server:
 Version:      17.03.1-ce
 API version:  1.27 (minimum version 1.12)
 Go version:   go1.7.5
 Git commit:   c6d412e
 Built:        Fri Mar 24 00:00:50 2017
 OS/Arch:      linux/amd64
 Experimental: true

[nomadme]

The same issue, having the port changed to 2377, although the generated command after initializing the swarm manager indicates part 2376.

[mdlinville]

Gwen wants to work on this.

[mdlinville]

@nishanttotla can you help us understand the mismatch with the ports?

[AnwarYagoub]

It worked with me following the solution listed in this post which suggested to advertise your address without specifying the port

$ docker-machine ssh myvm1 "docker swarm init --advertise-addr 192.168.99.100"

[nishanttotla]

The link posted by @AnwarYagoub seems reasonable to me. That is the expected behavior.

[gggh000]

Someone can tell me how to get this crap working?

[darmen]

HI @gggh000
I faced the same issue and just used an IP from docker-machine ls without a port part meaning that it will use default (2377) port.

[mmoser18]

Nope - just leaving away the port does not help!
I did so and got the very same error:

C:\windows\system32>docker-machine ls
NAME    ACTIVE   DRIVER   STATE     URL                        SWARM   DOCKER     ERRORS
myvm1   -        hyperv   Running   tcp://192.168.1.102:2376           v18.09.2
myvm2   -        hyperv   Running   tcp://192.168.1.113:2376           v18.09.2

C:\windows\system32>docker-machine ssh myvm1 "docker swarm init --advertise-addr 192.168.1.102"
Swarm initialized: current node (19gmiq7vr8smcdejw0p2rr0np) is now a manager.

To add a worker to this swarm, run the following command:

    docker swarm join --token SWMTKN-1-58ibmapquculh9rd6zloubaik8avcbd4v5z8ns5906lxu7pn8b-50x2vb5oa58ti259gmvgjg4jy 192.168.1.102:2377

To add a manager to this swarm, run 'docker swarm join-token manager' and follow the instructions.

C:\windows\system32>docker-machine ssh myvm2 "docker swarm join --token SWMTKN-1-58ibmapquculh9rd6zloubaik8avcbd4v5z8ns5906lxu7pn8b-50x2vb5oa58ti259gmvgjg4jy 192.168.1.113"
Error response from daemon: rpc error: code = Unavailable desc = all SubConns are in TransientFailure, latest connection error: connection error: desc = "transport: Error while dialing dial tcp 192.168.1.113:2377: connect: connection refused"
exit status 1

C:\windows\system32> 

Any clue anyone how to get this working?

[mmoser18]

My error! I had misunderstood that I should take the machine's IP address as last argument here, but to join the swarm one needs to specify the swarm-manager's IP, i.e. as provided in the output from the swarm init command!

I.e. the last command must be:
docker-machine ssh myvm2 "docker swarm join --token SWMTKN-1-58ibmapquculh9rd6zloubaik8avcbd4v5z8ns5906lxu7pn8b-50x2vb5oa58ti259gmvgjg4jy 192.168.1.102"
and not ... 113!

Working now!

[ghost]

The tutorial desperately needs updated with the correct procedure. The text at https://docs.docker.com/get-started/part4/ still maintains, as a default procedure, we should specify port 2376 when initializing a swarm manager... Despite the short quip about default management ports, the tutorial instructions can be confusing and frustrating to new users. I suggest the example command be this:
docker-machine ssh myvm1 "docker swarm init --advertise-addr <myvm1 ip:2377>"

Thus the note following about management ports will be intuitive.