Hide Forgot
Description of problem: Currently the RBD CRUD is supported through cli end point API(api/v2/cluster/<fsid>/cli). This introduces unnecessary hassle of constructing and parsing raw CLI output. Also the cli API is blocking which hangs if something goes wrong. To overcome these issues, require an API for RBD CRUD which is asynchronous
after further review we won't be able to deliver this work in time for 2.2
https://github.com/ceph/calamari/tree/12.17-wip-rbd-api When I first looked at the PR upstream for adding this feature I thought more was done. turns out we'll need to add a new service as a peer to cthulhu because RBD operations are data operations and when the cluster isn't healthy we would effectively be blocking the code that needs to report it as unhealthy