Is there an existing issue for this?
What are you currently unable to do
Incus environments are very diverse, and so are the configuration changes that can break them. You can erroneously change an address or bind a PCI device that leads your server to not being reachable, or, if your Incus server is your laptop, mess with its display and other niceties.
What do you think would need to be added
I think a --rollback=<n> flag (defaulting to 30 if the flag is passed with no value) should be added to the client to rollback configuration changes if an explicit prompt is not typed within n seconds.
The way I’d see it implemented is that setting this flag would make the configuration change an Incus operation with an automatic cancellation after n seconds, which gets terminated when the client tells the server that the user confirmed the change (a new endpoint could be used for that).
Is there an existing issue for this?
What are you currently unable to do
Incus environments are very diverse, and so are the configuration changes that can break them. You can erroneously change an address or bind a PCI device that leads your server to not being reachable, or, if your Incus server is your laptop, mess with its display and other niceties.
What do you think would need to be added
I think a
--rollback=<n>flag (defaulting to30if the flag is passed with no value) should be added to the client to rollback configuration changes if an explicit prompt is not typed within n seconds.The way I’d see it implemented is that setting this flag would make the configuration change an Incus operation with an automatic cancellation after n seconds, which gets terminated when the client tells the server that the user confirmed the change (a new endpoint could be used for that).