Page: Machines
Originally came in via the Red Hat bugzilla:
https://bugzilla.redhat.com/show_bug.cgi?id=1673158
Problem:
Sysadmins who handle VMs needs a way to snapshot the current state of a VM, do some potentially damaging operations, and have the ability to roll back to a previously well working state of the machine.
Initial scope:
Edit: outdated. Replaced by https://github.com/cockpit-project/cockpit/issues/13010#issuecomment-545006152

Some quick explanation for the mockup:
Some years ago I tested also snapshot on vm with raw disks and was possible:
with "virsh snapshot-create-as" it created additional qcow disks file with base file the raw ones; if critical operation "fails" stop the vm delete the qcow file and change the vm xml to repoint to the base raw file (virsh snapshot-revert to external snapshot was not supported, I didn't retest recently); if was ok use "virsh blockcommit" for consolidate qcow in the base file raw.
If I'll have time and needed I'll retest it and I'll post the exact procedure with exact command parameters.
Looks great!
Just some small comments:
"Create Snapshot" button should have default (white) design instead of primary (blue) since we use default design for all "Create something" buttons (apart from disks).
Is there any specific reason to have "Rename" and "Edit description" functionalities separate instead of single "Edit"?
@skobyda good calls! I've made the button default and merged the _Rename_ and _Edit description_ menu items.
Also discussed the design with @Ranelim and he pointed out it would be weird to single out the current snapshot, since the snapshot doesn't reflect the current state of the system, since a rollback could possibly have happened years ago.

Implementation happening in https://github.com/cockpit-project/cockpit/pull/13100
I'm interested in this feature. Thank you for your work.
Will be really pleased when this feature is available :-)
+1
This is done.
Most helpful comment
I'm interested in this feature. Thank you for your work.