notifies :restart, "docker_container[...]" doesn't seem to work about docker

[bflad]

I'm presuming the init_type is upstart from your other ticket. I wonder if
its related to other Upstart issues I've seen where a full stop and start
is required? Does debug output show it calling the underlying service
restart?

On Tue, Jan 28, 2014 at 4:17 PM, Wes Morgan [email protected]:

I added a restart notification to the template resource that creates a
config file one of my containers relies on. The whole template &
docker_container resources look like this:

docker_container 'registry' do
detach true
port '5000:5000'
env ['SETTINGS_FLAVOR=prod']
volume ['/mnt/docker-registry-storage:/tmp/registry', '/mnt/docker-registry-config:/docker-registry/config']end
template "/mnt/docker-registry-config/config.yml" do
owner "root"
group "docker"
mode "664"
action :create
source "docker-registry-config.yml.erb"
variables :config => {
:loglevel => "debug",
:storage => "local",
:storage_path => "/tmp/registry",
:secret_key => Chef::EncryptedDataBagItem.load("secrets", "docker_registry")["secret_key"]
}
notifies :restart, "docker_container[registry]" # <--- what I addedend

After updating the template, I saw

Recipe: base::docker

• docker_container[registry] action restart

in the Chef output, but the changes didn't take effect until I manually
stopped and re-ran the registry container.

Reply to this email directly or view it on GitHubhttps://github.com//issues/50
.

Brian Flad
http://about.me/bflad

[cap10morgan]

Yes, upstart init_type.

I'll try it again tomorrow w/ debug level logging.

[bflad]

Is this still an issue? Thanks.

[JeanMertz]

I just ran into this today. Haven't investigated it yet, but just wanted to let you know that it stil seems to be an issue in the latest version of the cookbook.

[bflad]

Good to know. I'd try doing two separate notifies (stop and start) for debugging to see if they fire or not. Otherwise, I'd love if we could have a test case that could fail until we fix it.

[StFS]

I'm not sure if I'm running into the same problem but on my installation on CentOS, docker seems to be a little "borked" after installing via this recipe. On a clean CentOS server, installing docker only (no containers) I get an error stating that the docker daemon isn't running. It isn't until I do a full manual restart (can't remember if restart works or if I have to do a stop and then a start or even reboot the machine). Do you think this is the same issue or should I file a new one (in which I'll provide better analysis of what actually happens)?

[bflad]

This sounds similar to #162. I would love to see the docker logs, if a restart fixes it, and if anyone has any additional insight. Thanks!

[tduffield]

@cap10morgan is this still an issue?

[fxposter]

This is still an issue and it is fixed by removing "normal exit 0 CONT" in ubuntu 14.04 (and greater, I suppose)

[jonapich]

I have the same problem. When I use a notifies :restart on a container, the container ends up in Exit (137) state and isn't restarted. When using stop followed by start though, it works.

I'm using the Ubuntu 14.04 community image from AWS.

[gionn]

:restart with upstart is borked, but maybe I've a fix.

[gionn]

If you want to try PR and test, just copy the template of the referenced PR in your cookbook (where you use docker_container resources), in templates/ and then use something like:

docker_container 'mycontainer' do
    init_template 'docker-container.conf.erb'
    cookbook 'yourcookbook'
end