Dynamically provision Stateful Persistent Replicated Cluster-wide Fabric Volumes & Filesystems for Kubernetes that is provisioned from an optimized NVME SPDK backend data storage stack.
Are you reporting an issue with existing content?
On linked page it is an advice to use kubectl mayastor upgrade command with --set-args , but this argument is invalid.
Tried for both 2.7.0 and 2.7.1 kubectl-mayastor extension.
Application 'upgrade' revision d0a6618f4898 (v2.7.0+0)
2024-10-28T13:02:47.386848Z INFO upgrade_job: Validated all inputs
at k8s/upgrade/src/bin/upgrade-job/main.rs:64
2024-10-28T13:02:47.725892Z INFO upgrade_job::helm::upgrade: Verified that openebs helm chart release 'openebs' has version 4.0.1
at k8s/upgrade/src/bin/upgrade-job/helm/upgrade.rs:359
2024-10-28T13:02:47.915995Z INFO upgrade_job::upgrade::data_plane: Skipping data-plane upgrade: All data-plane Pods are already upgraded
at k8s/upgrade/src/bin/upgrade-job/upgrade/data_plane.rs:43
but io_engines are still running on 1,2 cpus as before any changes.
I had also did helm upgrade with:
io_engine:
coreList: [2,3]
and also no change.
I did also kill -9 io_engine pid and it was also recreated on 1,2 cpus.
So not sure if that setting is working at all, maybe should just keep defaults 1,2 and isolate these cpus.
It is --set IIRC but actually that isn't going to work unless is done during the install. For post-install it needs the upgrade changes which @niladrih is working on.
@niladrih can you confirm here?
Are you reporting an issue with existing content? On linked page it is an advice to use
kubectl mayastor upgrade
command with--set-args
, but this argument is invalid. Tried for both 2.7.0 and 2.7.1 kubectl-mayastor extension.https://openebs.io/docs/user-guides/replicated-storage-user-guide/replicated-pv-mayastor/additional-information/performance-tips
what is the correct command?
btw. I did:
and job completed:
but io_engines are still running on 1,2 cpus as before any changes.
I had also did helm upgrade with:
and also no change.
I did also kill -9 io_engine pid and it was also recreated on 1,2 cpus.
So not sure if that setting is working at all, maybe should just keep defaults 1,2 and isolate these cpus.