TASK [deploy : Failed connection to remote repo] *******************************
System info:
Ansible 2.0.2.0; Linux
Trellis at “Wrap my.cnf password in quotes”
Git repo git@bitbucket.org:/.git cannot be accessed. Please verify
the repository exists and you have SSH forwarding set up correctly.
More info:
Ok, i got it, for some reason, when this file http://s000.tinyupload.com/index.php?file_id=50417395264007997836 is commited in git it makes problem - its automaticaly modified on server, so i cant pull from bitbucket, because there is a diff. That error message is misleading. When its deleted there is no problem. My question is, is there some option to force pull from git on server, no matter if there are any changes in files on server? thx
Thanks @payter86, I actually did resolved the issue by reading the comment from @swalkinshaw in another post.
Basically, my local system’s ssh key is being forwarded with the ssh ForwardAgent, the issue what that web user wasn’t able to connect to my gitlab server, because of the old known host entry.
I cleared the entry and checked the connection for web user via “ssh -T git@gitlab.com” and it did worked.
I learned that all following 3 has to be able to connect to gitlab / github repo.
Local Machine
Vagrant Box
Production server (digital Ocean droplet) - Specifically the web user.
roles/deploy/tasks/update.yml -> clone project files task thinks it’s failing on first run (and subsequent).
However
/srv/www/example.com/shared/source now contains the content of the repository.
So why does Ansible think that the task is failing?
The task and fail response match current Trellis master:
- name: Clone project files
git:
repo: "{{ project_git_repo }}"
dest: "{{ project_source_path }}"
version: "{{ project_version }}"
accept_hostkey: "{{ project.repo_accept_hostkey | default(repo_accept_hostkey | default(true)) }}"
ignore_errors: true
no_log: true
register: git_clone
- name: Failed connection to remote repo
fail:
msg: |
Git repo {{ project.repo }} cannot be accessed. Please verify the repository exists and you have SSH forwarding set up correctly.
More info:
> https://roots.io/trellis/docs/deploys/#ssh-keys
> https://roots.io/trellis/docs/ssh-keys/#cloning-remote-repo-using-ssh-agent-forwarding
when: git_clone | failed
Previous troubleshooting:
Successfully run ssh -T git@bitbucket.org as Vagrant user (from guest machine), as myself from local machine and as web from remote server, as long as I have connected with ForwardAgent yes:
# ~/.ssh/config
Host webdems
Hostname IP_ADDRESS_REMOTE_SERVER
User web
IdentityFile ~/.ssh/id_atlassian
ForwardAgent yes
This would be an the response when successful, I believe:
logged in as mikeill.
You can use git or hg to connect to Bitbucket. Shell access is disabled.
Some debugging code for roles/deploy/tasks/update.yml got me this far. Will include it.
- name: See Who I Am
command: whoami
register: who_i_am
- debug: msg="{{ who_i_am.stdout }}"
- name: See If I can Connect
command: ssh -T git@bitbucket.org
register: git_ssh_test
- debug: msg="{{ git_ssh_test.stdout }}"
- debug: msg="{{ git_ssh_test.stderr }}"
- debug: msg="{{ project_git_repo }}"
- debug: msg="{{ project_version }}"
- debug: msg="{{ project_source_path }}"
- debug: msg="{{ project.repo }}"
Output is:
TASK [deploy : debug] **************************************************************************************************
web
ok: [ip_address_here_dude]
TASK [deploy : See If I can Connect] ***********************************************************************************
changed: [ip_address_here_dude]
TASK [deploy : debug] **************************************************************************************************
logged in as mikeill.
You can use git or hg to connect to Bitbucket. Shell access is disabled.
ok: [ip_address_here_dude]
TASK [deploy : debug] **************************************************************************************************
ok: [ip_address_here_dude]
TASK [deploy : debug] **************************************************************************************************
git@bitbucket.org:myaccount/myrepo.git
ok: [ip_address_here_dude]
TASK [deploy : debug] **************************************************************************************************
master
ok: [ip_address_here_dude]
TASK [deploy : debug] **************************************************************************************************
/srv/www/example.com/shared/source
ok: [ip_address_here_dude]
TASK [deploy : debug] **************************************************************************************************
git@bitbucket.org:myaccount/myrepo.git
Here’s hoping one of you guys is feeling generous with your time…
Possible clue. Logged into the server where I have cloned the repository and a submodule seems to have been modified:
On branch master
Your branch is up-to-date with 'origin/master'.
Changes not staged for commit:
(use "git add <file>..." to update what will be committed)
(use "git checkout -- <file>..." to discard changes in working directory)
(commit or discard the untracked or modified content in submodules)
modified: site/web/dp/a-repository-here (new commits, modified content)
no changes added to commit (use "git add" and/or "git commit -a")
My issue was that there is a submodule (i know, i know) referenced in the repository, to which I hadn’t pushed my latest changes.
@swalkinshaw I wonder if it would make sense to have more specific output when the task fails. Also someone above has said that the Vagrant user needs to be able to access the repository. Is this the case from your experience? I wouldn’t have thought so. Thanks.