openstack team mailing list archive

[Jose Idar <jose.idar@xxxxxxxxxxxxx>]

I've noticed in testing that only way reliably to delete a volume 
attachment is to keep sending delete requests to the volume attachments 
api until the request is accepted and either a 2XX or 404 is returned.

Is there any plan to include some kind of atomic scheduled delete that 
would alleviate the need to do this by accepting the request and 
scheduling a delete for when that delete becomes possible?
(I understand that it's currently within the purview of the instance to 
make an attached volume ready for detaching, but even then it seems like 
more of an involved process than it should be, considering that at some 
level the API has to communicate with the instance/hypervisor in order 
to decide if it should allow the request or not)