swarm mode and overlay networks issues

[Riziero]

Hello,

I am creating a swarm cluster and then an overlay network which I am then using to spin up services. It appears that by trying to constraint a service to run on a specific node (--constraint node.hostname==node3) then the overlay network does not work.

As in I can not ping a service by its name from other services in the same network.
As soon as I get rid of the node contraint everything goes back to being pingable again.

Anyone have any clue about this?

Thanks!

[gianarb]

Hello @Riziero can you copy paste your docker info and docker version output?
Thanks

[aboch]

@Riziero In addition to what @gianarb already asked, please also add the info about how many managers/workers, the service creation command you are using, the network creation command, and any suspicious docker daemon warn/err logs if any. Thanks.

[Riziero]

@gianarb @aboch here's some more info.

I have 1 leader and 2 worker nodes:

$ docker node ls
ID HOSTNAME STATUS AVAILABILITY MANAGER STATUS
6pth9qm7d82j2324oqpd3cif9 node3 Ready Active
990z5o48vtsrfgrqfao6b3vjp node2 Ready Active
bpu3bwt85gon8sn0say2tcqgz * node1 Ready Active Leader

Docker info from node2:

$ docker info
Containers: 1
Running: 1
Paused: 0
Stopped: 0
Images: 24
Server Version: 1.12.6
Storage Driver: aufs
Root Dir: /var/lib/docker/aufs
Backing Filesystem: extfs
Dirs: 33
Dirperm1 Supported: true
Logging Driver: json-file
Cgroup Driver: cgroupfs
Plugins:
Volume: local
Network: bridge host null overlay
Swarm: active
NodeID: 990z5o48vtsrfgrqfao6b3vjp
Is Manager: false
Node Address: 172.28.128.6
Runtimes: runc
Default Runtime: runc
Security Options: apparmor seccomp
Kernel Version: 4.2.0-42-generic
Operating System: Ubuntu 15.10
OSType: linux
Architecture: x86_64
CPUs: 1
Total Memory: 1.954 GiB
Name: node2
ID: TQ74:V2AL:7IKR:IYXW:V5HW:UUWE:BXW7:HL7E:3WQQ:LVDI:CXYP:E762
Docker Root Dir: /var/lib/docker
Debug Mode (client): false
Debug Mode (server): false
Registry: https://index.docker.io/v1/
WARNING: No swap limit support
Labels:
provider=generic
Insecure Registries:
127.0.0.0/8
`

Docker version from inside the nodes:

vagrant@node2:/buildbot$ docker -v
Docker version 1.12.6, build 78d1802

Now what I am using to service2 on node2 and service3 on node 3:
docker service create \ --network foo \ --name postgres2 \ --constraint node.hostname==node2 \ postgres:9.5

docher ssh node2
docker exec -i -t ID_OF_THE_CONTAINER_FOR_SERVICE_2 bash
ping postgres3

gives:

root@f93d8621a144:/# ping postgres3
PING postgres3 (10.0.4.5): 56 data bytes
92 bytes from f93d8621a144 (10.0.4.4): Destination Host Unreachable
92 bytes from f93d8621a144 (10.0.4.4): Destination Host Unreachable
92 bytes from f93d8621a144 (10.0.4.4): Destination Host Unreachable

Note how something strange happens with IP, going from .5 to .4

Thanks!

[Riziero]

Hope this can help a bit.

Same exact test on trusty64 works well. The fail was with wily.
EDIT: Note that docker version provisioned by docker-machine is 1.3 (was 1.12).

[hezor]

I have this exact same problem with Docker Swarm (1.12.6 running on three CentOS 7.3 Swarm Nodes). So @Riziero did you get this solved?

[thaJeztah]

Let me close this ticket for now, as it looks like it went stale.