[api] support per-isogroup storageclasses #98
Merged
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
Users running Kubernetes < 1.12, in which topology-aware volume scheduling was introduced, cannot use a single storageclass for all their statefulsets as there is no guarantee that the PV for a pod will be created in the zone the pod is pinned to. This can cause pods to be stuck in an unschedulable state. This allows users to specify a StorageClass per-statefulset (zone). This solves the above problem as a single StorageClass can be created with attributes to specify that PVs created with it will be in a given zone (see https://v1-11.docs.kubernetes.io/docs/concepts/storage/storage-classes/).
schallert
force-pushed
the
schallert/storageclass
branch
from
967f158 to
851a2d8
Compare
schallert
commented
Mar 5, 2019
| spec: | ||
| image: quay.io/m3db/m3dbnode:latest | ||
| replicationFactor: 3 | ||
| numberOfShards: 4 |
There was a problem hiding this comment.
Note to self: change to 1024 or larger
nerd0
reviewed
Mar 6, 2019
| @@ -188,6 +192,9 @@ func GenerateStatefulSet( | |||
| } else { | |||
| template := cluster.Spec.DataDirVolumeClaimTemplate.DeepCopy() | |||
| template.ObjectMeta.Name = _dataVolumeName | |||
| if sc := isolationGroup.StorageClassName; sc != "" { | |||
There was a problem hiding this comment.
Is there anyway to ensure that there is a storage class of that name within the template itself? Do that need to validated?
There was a problem hiding this comment.
I think for now it's better to let Kube throw the error when it tries to create the statefulset and doesn't find it, otherwise we'd have to add new RBAC permissions to read StorageClasses
There was a problem hiding this comment.
Fair, I wanted to ensure that it would be clear to user that a storage class doesn't exist.
nerd0
approved these changes
Mar 6, 2019
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in 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.
Users running Kubernetes < 1.12, in which topology-aware volume
scheduling was introduced, cannot use a single storageclass for all
their statefulsets as there is no guarantee that the PV for a pod will
be created in the zone the pod is pinned to. This can cause pods to be
stuck in an unschedulable state.
This allows users to specify a StorageClass per-statefulset (zone). This
solves the above problem as a single StorageClass can be created with
attributes to specify that PVs created with it will be in a given zone
(see
https://v1-11.docs.kubernetes.io/docs/concepts/storage/storage-classes/).
Fixes #92