Major This is an LVMS Bug Report:
Please make sure that you describe your storage configuration in detail. List all devices that you plan to work with for LVMS as well as any relevant machine configuration data to make it easier for an engineer to help out.
Description of problem:
LVM storage is degraded after SNO node reboot
Version-Release number of selected component (if applicable):
LVM operator version: 4.14.6
Steps to Reproduce:
1. Add additonal disk to the SNO node 2. Install LVM Storage operator 3. Configure LVMCluster object like below to automatically discover the disks: apiVersion: lvm.topolvm.io/v1alpha1 kind: LVMCluster metadata: ..... ..... spec: storage: deviceClasses: - fstype: xfs name: vg1 thinPoolConfig: name: thin-pool-1 overprovisionRatio: 10 sizePercent: 90 4. After successfull configuration, take SNO node reboot (The device name change may take a few reboots). 5. LVM Cluster is now degraded: apiVersion: lvm.topolvm.io/v1alpha1 kind: LVMCluster metadata: ..... ..... spec: storage: deviceClasses: - fstype: xfs name: vg1 thinPoolConfig: name: thin-pool-1 overprovisionRatio: 10 sizePercent: 90 status: deviceClassStatuses: - name: vg1 nodeStatus: - devices: - /dev/sdc node: 98-f2-b3-20-8c-4c reason: 'failed to create/extend volume group vg1: failed to create or extend volume group "vg1" using command ''/usr/sbin/vgextend vg1 /dev/sda'': exit status 5' status: Degraded state: Degraded
Actual results:
Can't use LVM storage the topolvm-node pod is in crashloopbackoff state.
Expected results:
The LVM storage cluster should work consistently even after node reboot.
Additional info:
Instead of using the device names to perform `vg` operations, `devicepaths` of disks can be used to make the storage configuration persistent even after reboot. Customer can't use the below snippet as configuration for `LVMcluster` as they need to know the disks ahead of time: ~~~ spec: storage: deviceClasses: - name: vg1 deviceSelector: paths: - /dev/disk/by-path/pci-0000:87:00.0-nvme-1 - /dev/disk/by-path/pci-0000:88:00.0-nvme-1 thinPoolConfig: name: thin-pool-1 sizePercent: 90 overprovisionRatio: 10 ~~~ A script something like this [1] can help to dynamically figure out additional disks and use the device paths for making them ready for LVM storage. [1] https://docs.openshift.com/container-platform/4.13/scalability_and_performance/recommended-performance-scale-practices/recommended-etcd-practices.html#move-etcd-different-disk_recommended-etcd-practices
