Code review comment for lp:~rackspace-titan/nova/instance_states

Revision history for this message
Brian Waldon (bcwaldon) wrote :

> > This is absolutely incredible. You've done a great job, here. One general
> > comment before the line-by-line stuff:
> >
> > Can we align the states with respect to tense? I don't think all of our
> states
> > need to end in ING or ED. What do you think?
>
> The original design had all vm_states and task_states in the same tenses. No
> INGs or EDs.
>
> task_states.SCHEDULE just didn't have the same effect on me as
> task_states.SCHEDULING but really it's not a huge difference in my mind. We're
> just a few seds away from changing everything so it's not difficult. Some of
> them get more confusing IMO without ING or ED:
>
> NETWORK vs NETWORKING
> PAUSE vs PAUSED
> STOP vs STOPPED
>
> Would you recommend unification of tenses or just removal of all ING/ED/tense?

Actually, I think it makes a lot more sense now. Tasks make sense to end in -ING and vm_states in -ED (or no suffix at all). There are a few specific states I want to point out:

task_states.SPAWN -> task_states.SPAWNING

vm_states.VERIFY_RESIZE just seems weird. How does RESIZE_VERIFICATION sound, or WAITING?

Can task_states.UNPAUSING go away in favor of RESUMING? This is more of a question than a suggestion, I can see why we might want to leave it.

What about these for images:
task_states.SNAPSHOTTING -> IMAGE_SNAPSHOT
task_states.BACKING_UP -> IMAGE_BACKUP

task_states.HARD_REBOOTING should go away until we support it. If you want to leave it, can you rename it to REBOOTING_HARD?

task_states.PASSWORD -> UPDATING_PASSWORD

task_states.STARTING -> BOOTING
'starting' can easily mean a million different things

I don't see any 'shutdown' states.

« Back to merge proposal