When I run ansible-playbook server.yml -e env=staging for the first time, I get:
TASK [wordpress-setup : Create database of sites] ******************************
System info:
Ansible 2.0.2.0; Linux
Trellis at "Fix #569 - Only skip subdomains for non-www domains"
---------------------------------------------------
'dict object' has no attribute u'fapps.korak.info'
fatal: [fapps.korak.info]: FAILED! => {"failed": true}
RUNNING HANDLER [fail2ban : restart fail2ban] **********************************
changed: [fapps.korak.info]
to retry, use: --limit @server.retry
PLAY RECAP *********************************************************************
fapps.korak.info : ok=78 changed=6 unreachable=0 failed=1
This is my vault.yml
part
vault_wordpress_sites:
fapps.korak.info:
env:
And this is my wordpress_sites.yml
wordpress_sites:
fapps.korak.info:
site_hosts:
- fapps.korak.info
Note that both site name and staging subdomain are equal.
And those files are group_vars/staging/vault.yml
and group_vars/staging/wordpress_sites.yml
?
Yes staging folder, everything else as stock, changed the domains in development and staging folders. Development vagrants up ok, but staging is giving me problems.
So based on notice I got in the Ansible process, I supposed it could be due to same domain == site name/key in dictionary, but I don’t understand Ansible that well.
Did you finally solved the issue @Ivan_Svaljek ? I’ve got the same problem…
What I did is: bought a domain to differentiate staging host from site name, replaced all instances of fapps.korak.info in group_vars.
Now site name is fapps.info and staging host is staging.fapps.info and it seems to go ok.
But now that I think of it, it also seems to have been some instance of host name or site name in some stray trellis file, so I suggest you do a grep (or grepwin) through your hole trellis folder for those two names.
But it seems subdomains should be reserved for staging mode, I guess we’re not supposed to have the site name equal staging host name (but this is my guess).
Thanks for your impressions on this @Ivan_Svaljek . I will take this into account and hopefully I’ll be able to solve the issue.