openstack team mailing list archive

[John Garbutt <John.Garbutt@xxxxxxxxxx>]

> From: Vishvananda Ishaya [mailto:vishvananda@xxxxxxxxx] 
> Keep in mind that we actually have three options
Good point, I forgot about resize.

I guess we have:
- Live Migrate (with/without block migrate)
- (Non-live) Migrate
- Resize

I guess the more general way of looking at this is having an "optional scheduler hint":
- user might want to live-migrate to a different availability zone
- user might want to resize to same availability zone
- admin might what to migrate (live or otherwise) to a specific server

> Yun actually suggested that resize/migrate be simplified to do the following instead of scping the file over:
> * snapshot to glance
> * boot new image from snapshot
> This would definitely simplify the code, unfortunately it could have billing/metering repercussions.
Forgot about that, I like that too. To get the same semantics, I guess it would be:
stop VM, snapshot VM, start new VM, destroy old VM, allow user to delete snapshot.
Rollback just resumes VM and deletes snapshot.

I guess you could have something like "this will use data" warning you get on mobiles, but something like:
"This operation involves you being billed for a snapshot. You may delete the snapshot once the operation has completed"

What do people think about the billing issues from this change?

Cheers,
John