Symlink issues with deploys

bedrock
1 
TASK: [deploy | Create shared symlinks] *************************************** 
failed: [104.131.107.162] => (item={'path': 'web/app/uploads', 'src': 'uploads'}) => {"failed": true, "item": {"path": "web/app/uploads", "src": "uploads"}, "path": "/srv/www/domain.com/releases/20150502152244/web/app/uploads", "state": "absent"}
msg: Error while linking: [Errno 2] No such file or directory

FATAL: all hosts have already failed -- aborting
2

Is this still happening? If so make sure you have least bedrock-ansible code. I vaguely remember some changes to do with this.

3

I got this error as well on my first deploy. Will try to update bedrock-ansible.

4

No, it is not. I cloned the latest version of bedrock ansible and it went away. Whatever changes you folks implemented, it fixed the issue.

5

I’m sorry, but I’ve tried updating bedrock-ansible now by cloning it and inserting my details into /group_vars/staging and /hosts/staging, but I still get the same fail. Any ideas on what I should try next to get my deploy working?

TASK: [deploy | Create shared symlinks] ***************************************
failed: [ip] => (item={'path': 'web/app/uploads', 'src': 'uploads'}) => {"failed": true, "item": {"path": "web/app/uploads", "src": "uploads"}, "path": "/srv/www/domain.com/releases/20150511170726/web/app/uploads", "state": "absent"}
msg: Error while linking: [Errno 2] No such file or directory

FATAL: all hosts have already failed -- aborting
6

I just ended up cloning https://github.com/roots/roots-example-project.com and redoing my details and copying over the old bedrock site into site and it worked fine.

1 Like
7

Just ran into this same issue. If anyone set’s up their project like the roots example project, make sure to add subtree: site to their group_vars/production and staging files.

2 Likes
8

I have a similar problem. I’ve just cloned the latest Bedrock-ansible. Everything works smoothly with the provisioning, but when I come to deploy I get:

fatal: [xx.xxx.xx.xxx] => One or more undefined variables: 'dict object' has no attribute 'subtree'

FATAL: all hosts have already failed -- aborting

If I add subtree: site to production and staging group_vars I get:

TASK: [deploy | Move project subtree into root folder] ************************ 
failed: [xx.xxx.xx.xxx] => {"changed": true, "cmd": "mv /srv/www/test.co.uk/releases/20150527074840/site/* /srv/www/test.co.uk/releases/20150527074840", "delta": "0:00:00.004921", "end": "2015-05-27 07:48:52.973108", "rc": 1, "start": "2015-05-27 07:48:52.968187", "warnings": []}
stderr: mv: cannot stat ‘/srv/www/test.co.uk/releases/20150527074840/site/*’: No such file or directory

FATAL: all hosts have already failed -- aborting
9

Okay so now I understand what a subtree is. I have my site and ansible directories in different git repos, so I don’t need to use subtrees at all. I’ve tried reverting roles/deploy.defaults/main.yml to the previous version to no avail.

10

I had a look at the commit history on github and checked out the previous versions of both roles/deploy/defaults/main.yml and roles/deploy/tasks/main.yml and now I’m up and running again.

11

Are you saying that the very latest commit on master was giving you that error?

https://github.com/roots/bedrock-ansible/commit/56045fc3a05adc5c4cccfc805ee61814a705ac19 this commit should have fixed it.

12

Yes.

Let me check again just to make sure I’ve not done something stupid.

13

So I cloned a fresh copy and it’s all working again now. I’ve no idea what went wrong before.

The previous clone must have been older than I realised.

Sorry for wasting your time!

1 Like
14

Going to close this thread since it was an actual bug that’s been fixed.