[SOLVED] local-lvm became inactive after joining a cluster with different disk structure (no such logical volume pve/data)

[templar]

I had a cluster with classical disk roles: one for Proxmox OS, another for VMs. It worked. Then I wanted to add a simple small node with single disk for DNS and Ngynx. I've installed Proxmox there, choosing to format disk as ZFS. It was installed and created "local" and "local-lvm" storage. Then I joined it to the cluster and immediately after that "local-lvm" on fresh node got question mark and became inactive. Trying to access it throws error no such logical volume pve/data (500)On older nodes "local-lvm" is active. I suppose, it's because this new node got /etc/pve/storage.cfg file from the cluster, and other nodes have another configuration of local-lvm. How can I restore local-lvm here? Is it possible without reinstall of this node?


vgs returns nothing

~# pvesm status
no such logical volume pve/data
Name             Type     Status           Total            Used       Available        %
h.stor           cifs     active      3300184880      3098254896       201929984   93.88%
local             dir     active       705113088             128       705112960    0.00%
local-lvm     lvmthin   inactive               0               0               0    0.00%
zpool1        zfspool   disabled               0               0               0      N/A
zpool2        zfspool   disabled               0               0               0      N/A
zpool3        zfspool   disabled               0               0               0      N/A

/etc/pve/storage.cfg:

...
lvmthin: local-lvm
        thinpool data
        vgname pve
        content rootdir,images
...

 

[leesteken]

Go to Datacenter > Storage and make sure to select for each storage only the nodes that have that storage. Proxmox assumes all the nodes are equal, unless specified otherwise: https://pve.proxmox.com/pve-docs/pve-admin-guide.html#chapter_pmxcfs

 

[templar]

Go to Datacenter > Storage and make sure to select for each storage only the nodes that have that storage.

I did, thanks. Now it exists only on older nodes.
How to recreate it now on the new node? Both GUI and pvesm add ask for Volume Group, which I supposedly, don't have on a node with single ZFS formatted disk.

 

[sb-jw]

If you don't use LVM on the new one, then of course you won't be able to create it anymore. You then have to insert ZFS storage on the new one and LVM on the old one.

 

[templar]

If you don't use LVM on the new one, then of course you won't be able to create it anymore.

Yes, you are right. I have overseen, it wasn't "local-lvm", but "local-zfs". I tested same install in VM and found it out.
Now, I (re)added ZFS "rpool/data" in the "Datacenter > Storage" and can use it again.
So, case is solved