Open richardcase opened 2 years ago
With the delete being implemented in #144 having unused images is more likely.
I think creating a command like flintlock cleanup
, we can add flags --no-images
, --no-snapshots
, and --no-something-else
and let Ops set up how and when the command is executed. That way they can hook up in a cronjob, or ansible/puppet run,or basically anything they want. The main benefit from the ops side, they can easily enable, disable, or change when and what to clean up. Like do cleanup on x1, y1, and z1 daily, do x2 and y2 right after backup is done, and run x3 on every Monday midnight.
I like the idea of providing the command as well and leaving it up to the user to schedule its execution.
Another option is a maintenance api
This issue is stale because it has been open 60 days with no activity.
How about an option in the capi object that selects how cleanup as handled? This could end up in the miroVM details and thus mark how quickly cleanup should happen.
I think the issue with the cleanup in the microvm detail, we need cleanup from leftovers, things we couldn't/didn't delete after vm was deleted, like a snapshot. At that point, we don't have the vmspec anymore only leftover garbage on the system. So it's more like a "garbage collector" rather than a "clean up".
(or i misunderstood the microvm details part)
This issue is stale because it has been open 60 days with no activity.
Still valid.
This issue was closed because it has been stalled for 365 days with no activity.
We need to decide how we handle unused images, network interfaces.
When a microvm is deleted the snapshots will be deleted. This may leave a container image unused. What should we do with these unused images? A couple of options: