Files
cozystack/packages/apps/vm-disk
Myasnikov Daniil f73729937c docs(vm-disk): record what immediate binding costs on a node-pinned class
Requesting immediate binding for every source, not just `upload`, is
deliberate: a VMDisk is a standalone object populated before any VM
consumes it, so on a WaitForFirstConsumer StorageClass a disk with no VM
attached would never populate at all.

What that costs was recorded nowhere an operator reads. On a node-pinned
class the CDI worker pod chooses the node, so the volume binds where the
importer landed rather than where the VM will run, and a VM constrained
elsewhere -- a nodeSelector, a request only one node satisfies, or the
Windows affinity from _cluster.scheduling.dedicatedNodesForWindowsVMs --
stays unschedulable against a node-affinity conflict. One disk is enough
for that; it is not a multi-disk-only case. storageClass is immutable,
so the exit is delete and recreate. The chart default `replicated` binds
Immediate regardless and does not move.

Also pinned in tests, in both directions the annotation is emitted: per
source type (blank, image, disk) so a regression to the old upload-only
condition reddens, and on the existing-DataVolume path, where `helm
upgrade` retrofits the annotation onto disks created before it was
unconditional -- inert for a volume that has already bound, and now a
recorded decision rather than an unasserted side effect.

Signed-off-by: Myasnikov Daniil <myasnikovdaniil2001@gmail.com>
Assisted-By: LLM
2026-09-05 14:51:05 +05:00
..
2025-09-18 16:27:38 +02:00
2024-10-16 18:35:30 +02:00
2025-09-24 12:32:37 +02:00

Virtual Machine Disk

A Virtual Machine Disk

storageClass and source are annotated as immutable in the chart schema — see docs/storage-immutability.md for the contract and which consumers enforce it. DataVolume.spec is immutable in CDI, so the chart also fails the release on an attempted storageClass/source edit of an existing disk rather than silently reusing the old spec; delete and recreate the disk to change either field.

Every disk requests immediate binding (cdi.kubevirt.io/storage.bind.immediate.requested), so its volume is provisioned when the disk is created rather than deferred to the first VM that consumes it. That is deliberate: a VMDisk is a standalone object with a lifecycle of its own, and on a WaitForFirstConsumer StorageClass a disk that no VM has attached yet would otherwise never populate at all. The consequence on a node-pinned class such as local is that the CDI worker pod picks the node, so the volume lands where that pod was scheduled rather than where the VM will run: a VM whose placement is constrained — a nodeSelector, a resource request only one node satisfies, or the Windows affinity applied by _cluster.scheduling.dedicatedNodesForWindowsVMs — can then stay unschedulable against a node-affinity conflict, and because storageClass is immutable the only exit is deleting and recreating the disk. The chart default replicated binds Immediate regardless and is unaffected. Node-pinned classes are the case to weigh before using them for VM disks.

Parameters

Common parameters

Name Description Type Value
source The source image location used to create a disk. object {}
source.image Use image by name from default collection. *object null
source.image.name Name of the image to use. string ""
source.upload Upload local image. *object null
source.http Download image from an HTTP source. *object null
source.http.url URL to download the image. string ""
source.disk Clone an existing vm-disk. *object null
source.disk.name Name of the vm-disk to clone. string ""
optical Defines if disk should be considered optical. bool false
storage The size of the disk allocated for the virtual machine. quantity 5Gi
storageClass StorageClass used to store the data. string replicated