postgres_user.present fails but reports success

[olliewalsh]

This reports success:

auser:
  postgres_user.present:
    - superuser: true
    - createdb: true
    - password: apassword
    - user: postgres

When adding the user has failed with:
ERROR: conflicting or redundant options

[basepi]

Thanks for the report. We'll look into this.

[cachedout]

I cannot replicate this issue. With a new Postgres installation on Ubuntu with the SLS that you've posted, I get the following output from a state.sls run:

[INFO    ] Executing command '/usr/bin/psql --no-align --no-readline --dbname postgres -c \'CREATE USER "auser" WITH PASSWORD \'"\'"\'apassword\'"\'"\' CREATEDB SUPERUSER\'' as user 'postgres' in directory '/var/lib/postgresql'
[INFO    ] {'auser': 'Present'}
local:
----------
    State: - postgres_user
    Name:      auser
    Function:  present
        Result:    True
        Comment:   The user auser has been created
        Changes:   
                   ----------
                   auser:
                       Present

Summary
------------
Succeeded: 1
Failed:    0
------------
Total:     1

And then using psql:

postgres=# \du
                             List of roles
 Role name |                   Attributes                   | Member of 
-----------+------------------------------------------------+-----------
 auser     | Superuser, Create DB, Replication              | {}
 postgres  | Superuser, Create role, Create DB, Replication | {}

Could you please post a salt --versions-report and provide any additional details that you think might assist in reproducing this issue? Thanks!

[olliewalsh]

Ah, I missed an option. It fails silently when createuser is set too:

auser:
  postgres_user.present:
    - superuser: true
    - createdb: true
    - createuser: true
    - password: apassword
    - user: postgres

[cachedout]

Ah, that makes a lot more sense. That's an easy fix.

[cachedout]

This should be all fixed. @olliewalsh, could you please take a peek when you get a chance and see if this resolves the issue for you? If it does, please close the issue so we know that's it's all fixed. Thanks!

[cachedout]

I'm fairly confident this has been fixed in the recent work on the Postgres module. Since we've had no reply, I'm setting to closed for now. This module is under fairly heavy unit testing now, as well, so hopefully future regressions can be avoided. If this issue is still biting anybody, please feel free to re-open and we'll address it.