PV reservation with Local PV ZFS? #3819
Unanswered
fabriziogiudici
asked this question in
General
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
Hello. I'm working to evaluate OpenEBS as an improved way of managing volumes in a Kubernetes project that presently uses native storage supplied by the cloud provider. My first step is to integrate Local PV ZFS; when and if it will be consolidated I'll try Replicated PV Mayastor.
I followed the instructions and had the system up and running with no big troubles, but I have two issues that so far I wasn't able to fix:
pvc-__UUID__and I'd like to have a custom name, as it happens in my current system.reclaimPolicy: Retain), but if I undeploy/redeploy a package that is using a volume (or if I rebuild the system from scratch) new empty persistent volumes are created each time; the old ones are still in the zpool, but unused.This to me is really a problem since I can't resync the volumes from a backup every time I re-create the cluster, as it would take several minutes. Of course in production undeploying / rebuilding the system is something that I seldom do, but in the test environment it might happen a few times in a day when there is intense development.
Below the configuration files I'm using.
This is the StorageClass I've deployed:
This is the PVC:
I've tried to add a
volumeName: portainerwhich, according to the specs, should set the created volume, but the pvc just hangs inPendingstate.I've tried to configure the thing with the usual approach described at https://kubernetes.io/docs/concepts/storage/persistent-volumes/#reserving-a-persistentvolume:
and having an explicit PV (as documented also at https://kubernetes.io/blog/2019/01/15/container-storage-interface-ga/#pre-provisioned-volumes):
A pv with the desided name appears and it seems bound to the claim:
Also the pvc seems ok:
But the pod can't mount it:
And actually it doesn't appear in the
zfs liston the node where the zpool is accessible.I've read https://github.com/openebs/openebs/issues/479, but the proposed solution is cumbersome and not easily automatable. I have playbooks capable to create the infrastructure and deploy k8s from scratch without operator intervention and I want them to stay.
Please:
Thanks.
Beta Was this translation helpful? Give feedback.
All reactions