Bug 1526605

Description of problem:

https://access.redhat.com/documentation/en-us/red_hat_ceph_storage/3/html-single/administration_guide/#installing-the-red-hat-ceph-storage-dashboard


Following the documentation (above) to install ceph-grafana, skipping step 4 (not manually creating var.yml), it is failing with the following error:


ASK [ceph-grafana : Create Graphite DB tables] *************************************************************************************************************************************************************************************************************************
fatal: [ceph30]: FAILED! => {"failed": true, "msg": "The conditional check 'migrate_cmd.rc != 0 and 'table \"django_content_type\" already exists' not in migrate_cmd.stderr' failed. The error was: error while evaluating conditional (migrate_cmd.rc != 0 and 'table \"django_content_type\" already exists' not in migrate_cmd.stderr): Unable to look up a name or access an attribute in template string ({% if migrate_cmd.rc != 0 and 'table \"django_content_type\" already exists' not in migrate_cmd.stderr %} True {% else %} False {% endif %}).\nMake sure your variable name does not contain invalid characters like '-': argument of type 'StrictUndefined' is not iterable"}

If the customer uses the root user the installation completes as expected. However, called out that it's strange that the Ceph cluster itself deployed/installed properly when running the playbook as the ansible user, however when installing the dashboard it failed (until run as root).


How reproducible:
Following the steps as documented (and skipping optional step #4) - Always