Description of problem: If the user decides to change the cluster name to a non-default other than 'ceph' and is unfamiliar with the $CEPH_CONF environmental variable, all of the ceph commands in the cluster refuse to work as they're looking for a ceph config under /etc/ceph/ceph.conf, while the ceph-ansible playbook appropriately names the config after the cluster name. Version-Release number of selected component (if applicable): ceph-ansible-1.0.5-46.el7scon.noarch How reproducible: Always Steps to Reproduce: 1. Change the 'cluster:' line from 'ceph' to something else in group_vars/all 2. Run the ceph-ansible playbook Actual results: ceph commands do not work on cluster creation, stating no config found; users new to ceph may not understand what's occurring and may believe they're clusters are broken. Expected results: Alert users to utilize $CEPH_CONF var Additional info: While it'd be nice to set the env variable for users automatically, we'd probably only be able to set it for the root/ceph users and future users that are created would all need the variable to be added to their shell profiles in order to function. I feel the better way to handle this is to alert the user of the option on playbook completion, or if they're using RHSC within the web UI that a 'non-default' cluster name was used and point them at the documentation containing the $CEPH_CONF var.
(In reply to Kyle Squizzato from comment #0) > Additional info: > While it'd be nice to set the env variable for users automatically, we'd > probably only be able to set it for the root/ceph users and future users > that are created would all need the variable to be added to their shell > profiles in order to function. It's worth noting we could use /etc/profile here as well, but I'm a bigger fan of the "let the user handle it" approach.
Even so what you're saying is right this is difficult to achieve in Ansible. We cannot write an "info" message at the end of the play but only during the execution. This means the "info" message will be drowned by all the others messages. The only thing we can do is to write a doc section about this and put a comment in the code right after "cluster" in group_vars/all.yml. What do you think Kyle?
See: https://github.com/ceph/ceph-ansible/pull/1212
This should be in the next release after the 2.1 series.
Added doc text, let me know if that works for you.
LGTM :)
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2017:1496