PVC resizing support #616
PVC resizing support #616
Conversation
makhov
force-pushed
the
pvc-update
branch
10 times, most recently
from
01405de to
19b279a
Compare
makhov
force-pushed
the
pvc-update
branch
5 times, most recently
from
90e5907 to
5bc9c45
Compare
config/rbac/role.yaml
Outdated
| verbs: | ||
| - get | ||
| - list | ||
| - patch |
There was a problem hiding this comment.
I don't think we need to patch storageclasses objects 😄
| Env: []v1.EnvVar{ | ||
| {Name: "SVC_NAME", Value: kmc.GetEtcdServiceName()}, | ||
| }, | ||
| }, |
There was a problem hiding this comment.
What's the purpose of this? I get that it checks the etcd svc can resolve, but why do we need it?
There was a problem hiding this comment.
The check ensures the node is accessible from the other nodes during the join process.
| return fmt.Errorf("failed to delete pod '%s' for resizing: %w", fmt.Sprintf("%s-%d", stsName, i), err) | ||
| } | ||
| } else { | ||
| // Do not check other PVCs if expansion is not allowed |
There was a problem hiding this comment.
hmm, maybe we should somehow tell user this? I see two options, update the reconcile state and/or create an event
|
|
||
| return "200Mi" == sts.Spec.VolumeClaimTemplates[0].Spec.Resources.Requests.Storage().String(), nil | ||
| }) | ||
| s.Require().NoError(err) |
There was a problem hiding this comment.
I guess in the current test setup we cannot check much more as the local provisioners do not really support resizing?
In optimal case I think we'd need to verify couple things:
- resize actually happened
- no data has been lost
- child cluster still behaves properly
There was a problem hiding this comment.
Thanks! Updated the test.
makhov
force-pushed
the
pvc-update
branch
9 times, most recently
from
dc9aa9b to
df144a4
Compare
Signed-off-by: Alexey Makhov <[email protected]>
makhov
force-pushed
the
pvc-update
branch
2 times, most recently
from
df8e4bc to
088de56
Compare
| @@ -299,6 +299,18 @@ func (r *ClusterReconciler) initialCluster(kmc *km.Cluster, replicas int32) stri | |||
|
|
|||
| func (r *ClusterReconciler) generateEtcdInitContainers(kmc *km.Cluster) []v1.Container { | |||
| return []v1.Container{ | |||
| { | |||
| // Wait for the pods dns name is resolvable, since it takes some tima after the pod is created | |||
There was a problem hiding this comment.
There is a typo here, tima instead of time
| return false, nil | ||
| } | ||
|
|
||
| return "250Mi" == sts.Spec.VolumeClaimTemplates[0].Spec.Resources.Requests.Storage().String(), nil |
There was a problem hiding this comment.
I think we should also verify that the etcd volume has been resized to 70Mi
There was a problem hiding this comment.
Yeah, makes a lot of sense, thanks! Added the check for etcd statefulset
Signed-off-by: Alexey Makhov <[email protected]>
Fixes #610
k0smotron removes and recreates the Statefulset with an updated size. If the StorageClass allows expansion, k0smotron resizes PVCs and deletes pods to apply the changed PVC.
Also includes a minor etcd improvement, that makes check-etcd-scaling test pass