Ceph » devops

To improve the knowledge (and debugging) about what ceph-deploying is doing with ceph behind the scenes we should be outputting the actual remote command when performing an action.

Because the compiled functions are rather big, they usually contain a few of the subprocess calls within, which is convenient but problematic to be able to achieve an "display what is going on for every remote action".

One way to help with this is to create a 'backend' for remote actions. For example, lets take `mon create` that performs the following remote actions:

• if path doesn't exist create it
• if done path doesn't exist create it * if tmp path doesn't exist create it * write the monitor keyring * call ceph-mon --mkfs * unlink keyring * create done file
• if init file doesn't exist, create it
• start the ceph-mon daemon

That means we have 3 items at the very minimum and 9 at the most, but all of them could be smaller, contained, testable, and easier to tell the user what is some command about to do.

If they are all into one single function (as it is right now) this is non-trivial.