Error: no available IPv6 addresses on this network's address pools

[alpo]

Output of docker version:

Client:
 Version:      1.9.1
 API version:  1.21
 Go version:   go1.4.3
 Git commit:   9894698
 Built:        
 OS/Arch:      linux/amd64

Server:
 Version:      1.9.1
 API version:  1.21
 Go version:   go1.4.3
 Git commit:   9894698
 Built:        
 OS/Arch:      linux/amd64

Output of docker info:

Containers: 222
Images: 36
Server Version: 1.9.1
Storage Driver: overlay
 Backing Filesystem: extfs
Execution Driver: native-0.2
Logging Driver: json-file
Kernel Version: 4.3.6-coreos
Operating System: CoreOS 899.15.0
CPUs: 16
Total Memory: 31.47 GiB
Name: something
ID: 5GNN:MBJE:VTQ6:AUVN:CWMI:FKEH:NWZ7:C3OA:3GST:6E6A:RCD4:NW3V

Additional environment details (AWS, VirtualBox, physical, etc.):

physical

DOCKER_OPTS=--ipv6=true --fixed-cidr-v6=something/80 --userland-proxy=false

Unit files (nothing special)

[Unit]
Description=blah
After=docker.service
Requires=docker.service
[Service]
TimeoutStartSec=0
RestartSec=60
Restart=always
ExecStartPre=-/usr/bin/docker kill blah1
ExecStartPre=-/usr/bin/docker rm blah1
ExecStart=/usr/bin/docker run --name blah1 --dns blabla --publish blabla ownimages/blablabla
ExecStop=/usr/bin/docker stop blah1
[Install]
WantedBy=agents.target

Steps to reproduce the issue:

1. Create ~200 containers with launch via systemd units.
2. Launch containers

Describe the results you received:

Part of containers failed to launch with error message

Error: no available IPv6 addresses on this network's address pools

Describe the results you expected:

Successful launch

Additional information you deem important (e.g. issue happens only occasionally):

docker ps hangs. After command sudo systemctl restart docker the docker daemon becomes a zombie process and only reboot helps to restore docker 'working'.

docker network inspect bridge sometime works, sometime gives error: "An error occurred trying to connect: Get http:///var/run/docker.sock/v1.21/networks/bridge: EOFnull"

[aboch]

Likely a duplicate of #18527

[thaJeztah]

@alpo this issue should be resolved in docker 1.10.x and above. Can you test if you're still able to reproduce this on docker 1.10 or 1.11?

[alpo]

@thaJeztah After migration to CentOS 7 with docker 1.9.1 and CoreOS alpha 1010.1.0 with docker 1.10.3 no more weird messages and deadlocks in both cases.
PS: CoreOS alpha setup is same, but in the part migrated to CentOS IPv6 in docker is not configured, IPv6-demanding applications moved out of containers.

[mavenugo]

Closing this based on the above comment.