Change recovery behavior of NOT_FOUND in ControllerUnpublishVolume to a SHOULD #383
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
jdef
approved these changes
Aug 15, 2019
saad-ali
approved these changes
Aug 16, 2019
There was a problem hiding this comment.
This seems reasonable until we can fix all of this in CSI 2.0 (#382).
/lgtm
/approve
|
ping @julian-hj @ddebroy @jieyu |
julian-hj
approved these changes
Aug 27, 2019
jieyu
approved these changes
Aug 28, 2019
# for free
to join this conversation on GitHub.
Already have an account?
# to comment
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Outdated suggestions cannot be applied.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.
/assign @saad-ali @jdef @jieyu
follow-up to: #375
The reasoning why we want to change this to a
SHOULDis because any caller has no actual way of verifying that the volume exists without "asking" the driver. There is currently no call to find that exceptListVolumeswhich is an optional capability. Therefore, aMUSTto verify the volume exist after this error is actually logically impossible.We relax this to
SHOULDand this actually works really well with the aboveOKcode return if the driver already knows the volume/node is missing and that means the volume isControllerUnpublished.This is a non-breaking change to move the Spec into possibility of real compliance; however, in reality the Kubernetes external attacher already does not implement the recovery behaviors (for the above reasons). A proposal for making these changes less vague and more concrete (but breaking) is here: #382