Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
allow instance_reconfigure to resize a stopped instance #6774
base: main
Are you sure you want to change the base?
allow instance_reconfigure to resize a stopped instance #6774
Changes from 2 commits
0159e95
60afb8c
fd7e5a1
e76814f
57087c4
6217757
5c3efb0
a6d7544
File filter
Filter by extension
Conversations
Jump to
There are no files selected for viewing
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It occurs to me that
OK_TO_SET_SIZE_STATES
andOK_TO_SET_BOOT_DISK_STATES
are the same list of states, and what we really want in both cases is "states in which there is no VMM actively incarnating this instance". I kind of wonder if, rather than duplicating in both the instance-resize and instance-set-boot-disk queries, we should just have a constant onInstanceState
that's likeNOT_INCARNATED_STATES
or similar, so that these queries and any future ones can all use the same set of states, and if we add new states that have "not incarnated" semantics, we can add them to the one definition of that list of states, instead.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
on one hand i agree (and enough to go just do the thing, there's now
InstanceState::NO_VMM_STATES
*), but from the conversations we've had we've been pretty close to having caveouts for specific not-incarnated states and fields. namely "is it OK to set the boot disk inCreating
", and does allowing that allow correctness issues (no) or a possibly-confusing-but-correct API result (yes). otoh "actions we can take on not-incarnated instances" is a much easier concept than having to think through each kind of action and state combination, so it's nicer in an axiomatic way.*: after writing this i realize why you said
NOT_INCARNATED
rather thanNO_VMM
- because there is aNoVmm
state which is very distinct from what this list is getting at. so i should go change that name.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah, my thinking is that we should avoid having weird exceptions for setting specific parameters unless it's absolutely necessary. I'd be quite surprised if we don't basically just end up being able to divide every possible
instance_reconfigure
into either "can do this only if there is no Real Life VMM Process incarnating the instance" and "can do this whenever", and I'd prefer to have to burn that bridge only once we come to it...