Plugin not displaying all block devices

[greg80303]

Two of the nodes in my cluster have 6, 2TB drives. For some reason the plugin is only showing the block devices for 4 of the drives (always omitting the first and last drive:

$ kubectl mayastor get block-devices den-w-jhf6h04

 DEVNAME   DEVTYPE  SIZE    AVAILABLE  MODEL             DEVPATH                                                                            MAJOR  MINOR  DEVLINKS 
 /dev/sdb  disk     1.7TiB  yes        Dell_Ent_NVMe_CM  /devices/pci0000:64/0000:64:04.0/0000:65:00.0/host2/target2:2:1/2:2:1:0/block/sdb  8      16     "/dev/disk/by-id/scsi-SNVMe_Dell_Ent_NVMe_CM_33M0A0WZTCD8", "/dev/disk/by-id/scsi-201546a1be28ee38c", "/dev/disk/by-path/pci-0000:65:00.0-scsi-0:2:1:0", "/dev/disk/by-diskseq/10" 
 /dev/sdc  disk     1.7TiB  yes        Dell_Ent_NVMe_CM  /devices/pci0000:64/0000:64:04.0/0000:65:00.0/host2/target2:2:2/2:2:2:0/block/sdc  8      32     "/dev/disk/by-id/scsi-SNVMe_Dell_Ent_NVMe_CM_33M0A0VYTCD8", "/dev/disk/by-diskseq/11", "/dev/disk/by-id/scsi-2012e6a1be28ee38c", "/dev/disk/by-path/pci-0000:65:00.0-scsi-0:2:2:0" 
 /dev/sdd  disk     1.7TiB  yes        Dell_Ent_NVMe_CM  /devices/pci0000:64/0000:64:04.0/0000:65:00.0/host2/target2:2:3/2:2:3:0/block/sdd  8      48     "/dev/disk/by-id/scsi-SNVMe_Dell_Ent_NVMe_CM_6DW0A1VU0NN3", "/dev/disk/by-diskseq/12", "/dev/disk/by-path/pci-0000:65:00.0-scsi-0:2:3:0", "/dev/disk/by-id/scsi-201bea31fe28ee38c" 
 /dev/sde  disk     1.7TiB  yes        Dell_Ent_NVMe_CM  /devices/pci0000:64/0000:64:04.0/0000:65:00.0/host2/target2:2:4/2:2:4:0/block/sde  8      64     "/dev/disk/by-path/pci-0000:65:00.0-scsi-0:2:4:0", "/dev/disk/by-id/scsi-SNVMe_Dell_Ent_NVMe_CM_6DW0A1VW0NN3", "/dev/disk/by-id/scsi-201c0a31fe28ee38c", "/dev/disk/by-diskseq/13"

If I list the block devices on the node itself:

den-w-jhf6h04:~$ lsblk
NAME   MAJ:MIN RM   SIZE RO TYPE MOUNTPOINTS
sda      8:0    0   1.7T  0 disk 
sdb      8:16   0   1.7T  0 disk 
sdc      8:32   0   1.7T  0 disk 
sdd      8:48   0   1.7T  0 disk 
sde      8:64   0   1.7T  0 disk 
sdf      8:80   0   1.7T  0 disk 
sdg      8:96   0 447.1G  0 disk 
├─sdg1   8:97   0     1G  0 part /boot/efi
└─sdg2   8:98   0   446G  0 part /var/lib/kubelet/pods/0c9dee03-296a-4b2c-b6fc-91cad4f7f11d/volumes/kubernetes.io~local-volume/pvc-57ffcc24-da62-4205-88b0-2d815c8c6f7d.....

Is there a limit for block devices on a node? Or is there something about these devices that are keeping them from being discovered by the control plane and plugin? I see this exact same behavior on both of the nodes that have this drive configuration.

[tiagolobocastro]

hmm could you please try kubectl mayastor get block-devices den-w-jhf6h04 -a ?

[tiagolobocastro]

Ah, it's because it has a filesystem already:
/dev/sdf disk 1.7TiB no ext4 5ed34d7d-e87d-44d1-9d06-acf075e98ddd

[greg80303]

These devices used to be part of a hardware RAID array, but I deleted the virtual disk and marked all the drives as non-RAID in the hardware controller. Anything else I need to do make them "available"?

[greg80303]

ok -- let me double-check each drive to make sure there are no partitions and no filesystems.

[tiagolobocastro]

You need to remove the filesystem

[greg80303]

Thank you. The --all argument helped a bunch