You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
DeleteSnapshot may be called while that snapshot is in use as a volume source in a concurrent CreateVolume call, creating a race condition.
COs should be instructed how to deal with various situations when these two calls interact. Specifically, CO must block all CreateVolume calls with a snapshot volume source while that snapshot is being deleted, and vice versa - it must block all DeleteSnapshot calls while that snapshot is being used as a volume source in a CreateVolume call.
Is this something that belongs to the CSI spec? The snapshot controller in Kubernetes already does this, and attempting to delete such snapshots defers their deletion until the CreateVolume call(s) in question finish processing the request. I think it would be best to make this (or similar) behavior official by having it stated explicitly in the spec.
The text was updated successfully, but these errors were encountered:
On Fri, Oct 8, 2021, 9:37 AM Róbert Vašek ***@***.***> wrote:
DeleteSnapshot may be called while that snapshot is in use as a volume
source in a concurrent CreateVolume call, creating a race condition.
COs should be instructed how to deal with various situations when these
two calls interact. Specifically, CO must block all CreateVolume calls
with a snapshot volume source while that snapshot is being deleted, and
vice versa - it must block all DeleteSnapshot calls while that snapshot
is being used as a volume source in a CreateVolume call.
Is this something that belongs to the CSI spec? The snapshot controller in
Kubernetes already does this, and attempting to delete such snapshots
defers their deletion until the CreateVolume call(s) in question finish
processing the request. I think it would be best to make this (or similar)
behavior official by having it stated explicitly in the spec.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#493>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAR5KLCFR6Y4REYP44KVTHLUF3XYNANCNFSM5FTUSEGA>
.
Triage notifications on the go with GitHub Mobile for iOS
<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>
or Android
<https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.
DeleteSnapshot
may be called while that snapshot is in use as a volume source in a concurrentCreateVolume
call, creating a race condition.COs should be instructed how to deal with various situations when these two calls interact. Specifically, CO must block all
CreateVolume
calls with a snapshot volume source while that snapshot is being deleted, and vice versa - it must block allDeleteSnapshot
calls while that snapshot is being used as a volume source in aCreateVolume
call.Is this something that belongs to the CSI spec? The snapshot controller in Kubernetes already does this, and attempting to delete such snapshots defers their deletion until the
CreateVolume
call(s) in question finish processing the request. I think it would be best to make this (or similar) behavior official by having it stated explicitly in the spec.The text was updated successfully, but these errors were encountered: