-
Notifications
You must be signed in to change notification settings - Fork 242
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
disk: attach disks with serial number at controller #1199
base: master
Are you sure you want to change the base?
Conversation
c1e5398
to
30d433a
Compare
/unhold |
30d433a
to
25da495
Compare
25da495
to
a0e9edf
Compare
a0e9edf
to
336bd7a
Compare
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: huww98 The full list of commands accepted by this bot can be found here.
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
336bd7a
to
72d6526
Compare
based on #1330 |
72d6526
to
8b269b9
Compare
057c8cb
to
5b05b41
Compare
Don't assume the serial is the volume ID without "d-" prefix.
Eliminate the risk of finding the wrong disk. It may take a while for the disk device to appear after the OpenAPI returns in_use. e.g. when the previous detach is blocked, the attach will also be blocked.
Attach and detach must be migrated at once, because NodeUnstageVolume may be followed immediately by another NodeStageVolume, without controller actions. So if we detach disk at NodeUnstageVolume, NodeStageVolume will not find the disk. Note: user must upgrade controller before node, to avoid any missing detach. This is the first step to move all attach/detach back to controller, in order to: - reduce RAM permission on node - more effective request batching - rely less on node, which can be down - resolve some rare racing issue
5b05b41
to
f598525
Compare
What type of PR is this?
/kind cleanup
What this PR does / why we need it:
Attach and detach must be migrated at once, because NodeUnstageVolume may be followed immediately by anoterh NodeStageVolume, without controller actions.
So if we detach disk at NodeUnstageVolume, NodeStageVolume will not find the disk.
Note: user must upgrade controller before node, to avoid any missing detach.
This is the first step to move all attach/detach back to controller, in order to:
Which issue(s) this PR fixes:
Fixes #
Special notes for your reviewer:
/hold
merge after #1087
Does this PR introduce a user-facing change?
Additional documentation e.g., KEPs (Kubernetes Enhancement Proposals), usage docs, etc.: