This repository has been archived by the owner. It is now read-only.
Docker daemon not getting options from flannel_docker_opts.env #1659
Assignees
Milestone
Comments
|
Is |
|
@crawford Is there anything special for enabling it? All I have is this piece in the - name: docker.service
command: start
drop-ins:
- name: 50-require-flannel.conf
content: |
[Unit]
Requires=flanneld.service
After=flanneld.serviceAlso, the same |
|
Enabling the service allows systemd to manage the startup (which usually results in that service starting before coreos-cloudinit). You can check to see if it's enabled with |
|
@crawford Here is the output |
|
Okay, that looks fine. I tried reproducing the issue with the config you provided and was unable to do so. I watched as systemd waited for Can you attach the output of |
|
@crawford Thanks. Here is the output: Just out of curiosity, did you test on Alpha 1221 version? |
|
I was actually able to reproduce the failure. |
|
To work around the issue, you can use the following config (I made flannel-docker-opts.service a oneshot): #cloud-config
coreos:
units:
- name: etcd2.service
command: start
- name: flannel-docker-opts.service
drop-ins:
- name: oneshot.conf
content: |
[Service]
Type=oneshot
Restart=no
- name: flanneld.service
command: start
drop-ins:
- name: 50-network-config.conf
content: |
[Service]
ExecStartPre=/usr/bin/etcdctl set /coreos.com/network/config '{ "Network": "10.100.0.0/16" }'
- name: docker.service
command: start
drop-ins:
- name: 50-require-flannel.conf
content: |
[Unit]
Requires=flanneld.service
After=flanneld.service |
|
@crawford Yup that worked. Would you mind explaining the |
|
|
|
@crawford Great. Thanks! |
|
@crawford Can you please consider re-opening this? I'm still able to re-produce this bug in Beta 1235.1.0. Please see this gist.
|
|
@zihaoyu docker needs to wait for |
|
|
Oh, I forgot you were using coreos-cloudinit. I don't have a good solution for you. Your two options are either switch to Ignition (preferred) or restructure your cloud-config to ensure the options are created. Something like this: #cloud-config
coreos:
units:
- name: etcd2.service
command: start
- name: flanneld.service
command: start
drop-ins:
- name: 50-network-config.conf
content: |
[Service]
ExecStartPre=/usr/bin/etcdctl set /coreos.com/network/config '{ "Network": "10.100.0.0/16" }'
- name: flannel-docker-opts.service
command: start
- name: docker.service
command: start
drop-ins:
- name: 50-require-flannel.conf
content: |
[Unit]
Requires=flanneld.service
After=flanneld.service |
|
@crawford OK I will check out Ignition. |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Issue Report
Bug
CoreOS Version
Environment
CoreOS on AWS.
Expected Behavior
Docker daemon gets the
--bipand other command line options that are generated by flanneld and are available in/run/flannel/flannel_docker_opts.env.Actual Behavior
Docker daemon starts without
--bipand other command line flags. Below is mycloud-configYAML passed asuser-data:And I can see the generated
flannel_docker_opts.env:However, docker daemon starts without these flags:
Interestingly, if I restart
docker.service, Docker daemon is able to pick up the flags:I wonder if there is a problem in the order in which services are started by systemd? I looked through the unit files of
flanneld.service,flannel-docker-opts.service, anddocker.service, they all seem to be in correct order:flanneld->flannel-docker-opts->docker.Reproduction Steps
user-data.Other Information
CoreOS Stable and Beta AMIs work correctly at the moment. Alpha is broken. I noticed that Alpha ships with Docker 1.12, wondering if that's the problem.
The text was updated successfully, but these errors were encountered: