Error adding Windows node to Linux Swarm Cluster #30051

Closed
fabiendibot opened this Issue Jan 11, 2017 · 5 comments

Projects

None yet

4 participants

@fabiendibot
fabiendibot commented Jan 11, 2017 edited

Description

When creating a Swarm mode cluster on Linux with Docker.1.13.0-rc5 and adding a Windows 2016 agent running the same docker version, the windows agent is joining the cluster, but there are errors on Windows deamon.

time="2017-01-11T09:57:10.068352700+01:00" level=info msg="API listen on //./pipe/docker_engine"
time="2017-01-11T09:57:10.068352700+01:00" level=info msg="Initializing Libnetwork Agent Listen-Addr=0.0.0.0 Local-addr=
172.16.15.190 Adv-addr=172.16.15.190 Remote-addr =172.16.15.40"
time="2017-01-11T09:57:10.068352700+01:00" level=info msg="Gossip cluster hostname WIN-TQUGBJP8UUL-42912db34b6c"
time="2017-01-11T09:57:10.516373900+01:00" level=info msg="HNSNetwork Request ={\"Name\":\"vzsisuzwspe4dwcykg2z5d5ex\",\
"Type\":\"overlay\",\"Subnets\":[{\"AddressPrefix\":\"10.255.0.0/16\",\"GatewayAddress\":\"10.255.0.1\",\"Policies\":[{\
"Type\":\"VSID\",\"VSID\":4096}]}]}"
time="2017-01-11T09:57:10.525373900+01:00" level=error msg="Failed creating ingress network: HNS failed with error : Cat
astrophic failure "
time="2017-01-11T09:57:10.787392500+01:00" level=warning msg="memberlist: Got ping for unexpected node 'WIN-TQUGBJP8UUL-
b39e35da08c1' from=172.16.15.42:7946"

I can do some actions like adding labels to the node, bu when i triy to deploy a stack using constraints on the Windows agent i got this as well.

"Type\":\"overlay\",\"Subnets\":[{\"AddressPrefix\":\"10.0.0.0/24\",\"GatewayAddress\":\"10.0.0.1\",\"Policies\":[{\"Typ
e\":\"VSID\",\"VSID\":4097}]}]}"
time="2017-01-11T07:29:36.884146000+01:00" level=error msg="fatal task error" error="HNS failed with error : Catastrophi
 c failure " module="node/agent/taskmanager" task.id=l16wlt3w7579s8ae0xhbzz0wx
 time="2017-01-11T07:29:37.317158400+01:00" level=error msg="failed to deactivate service binding for container demo_sql.
1.v6faq67j4toaubzxsqko3qefs" error="No such container: demo_sql.1.v6faq67j4toaubzxsqko3qefs" module="node/agent"
 time="2017-01-11T07:29:41.877276000+01:00" level=info msg="HNSNetwork Request ={\"Name\":\"357brmk6puvawm18cx9jumpfm\",\
"Type\":\"overlay\",\"Subnets\":[{\"AddressPrefix\":\"10.0.0.0/24\",\"GatewayAddress\":\"10.0.0.1\",\"Policies\":[{\"Typ
e\":\"VSID\",\"VSID\":4097}]}]}"
 time="2017-01-11T07:29:42.328290400+01:00" level=error msg="failed to deactivate service binding for container demo_sql.
1.c15frilaxzt1t6h34dsnnymnr" error="No such container: demo_sql.1.c15frilaxzt1t6h34dsnnymnr" module="node/agent"
time="2017-01-11T07:29:52.345532300+01:00" level=error msg="fatal task error" error="HNS failed with error : Catastrophi
c failure " module="node/agent/taskmanager" task.id=nu9d3zxcqqcetq8ozg2ozzui9

Steps to reproduce the issue:

  1. Build a Swarm mode cluster using Linux
  2. Add a Windows 2016 Datacenter agent

Describe the results you received:

I can't use the windows node in a Swarm mode cluster

Describe the results you expected:

A Windows/Linux Swarm mode cluster :)

Output of docker version:

PS C:\Users\Administrator> docker version
Client:
 Version:      1.13.0-rc5
 API version:  1.25
 Go version:   go1.7.3
 Git commit:   43cc971
 Built:        Thu Jan  5 03:07:30 2017
 OS/Arch:      windows/amd64

Server:
 Version:      1.13.0-rc5
 API version:  1.25 (minimum version 1.24)
 Go version:   go1.7.3
 Git commit:   43cc971
 Built:        Thu Jan  5 03:07:30 2017
 OS/Arch:      windows/amd64
 Experimental: false

Output of docker info:

PS C:\Users\Administrator> docker info
Containers: 0
 Running: 0
 Paused: 0
 Stopped: 0
Images: 1
Server Version: 1.13.0-rc5
Storage Driver: windowsfilter
 Windows:
Logging Driver: json-file
Plugins:
 Volume: local
 Network: l2bridge l2tunnel nat null overlay transparent
Swarm: active
 NodeID: valri8mwlxaujzkbr4wsnhts1
 Is Manager: false
 Node Address: 172.16.15.190
 Manager Addresses:
  172.16.15.40:2377
  172.16.15.41:2377
  172.16.15.42:2377
Default Isolation: process
Kernel Version: 10.0 14393 (14393.693.amd64fre.rs1_release.161220-1747)
Operating System: Windows Server 2016 Datacenter
OSType: windows
Architecture: x86_64
CPUs: 1
Total Memory: 4 GiB
Name: WIN-TQUGBJP8UUL
ID: TQHZ:LRRZ:EQ5P:IYVW:5SW4:QHCM:VDBO:SBTT:E3V6:XXTK:KCPC:F446
Docker Root Dir: C:\ProgramData\docker
Debug Mode (client): false
Debug Mode (server): false
Registry: https://index.docker.io/v1/
Experimental: false
Insecure Registries:
 127.0.0.0/8
Live Restore Enabled: false

Maybe @msabansal or @jhowardmsft have an idea ?

@PatrickLang

Ping @Kallie-b and @jmesser81 on the HNS error. I think a Windows change is still needed for overlay networking to work.

@kallie-b

I believe that is correct, but I'm a little unsure of how this works if the Swarm manager is a Linux node. Will check on this and report back.

@kallie-b

Apologies for the delay--

Yes, overlay networking is not yet supported on Windows.

For a Windows host to be added to a Swarm cluster (whether it is a worker or manager node in the swarm), overlay is required. Currently the overlay driver is not supported on Windows, so multi-node Swarm scenarios are not yet enabled. As far as cross-OS support, we know that it's important and we're currently investigating what it might take to achieve this :)

@fabiendibot

Thx @kallie-b for the information, if you need me to do some tests, i've signed a MVP NDA, so let me know :)

@fabiendibot fabiendibot reopened this Jan 14, 2017
@kallie-b

Oh awesome, that's good to know :) I'll keep you in mind and let you know if an opportunity arises.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment