Joining a new member to a cluster with a linstor storage pool fails

[hosaka]

Initially posted this in lxc/incus-os#1266

The setup

• Cluster provisioned by the Operations Center Version: 0.8.1
• 3 existing members, each with the incus-linstor application installed and the linstor service enabled
• New member is a 4th server intended to be a diskless linstor satellite
• Linstor controller in a Debian container (project internal with meshbr0 network) on member node-1
• IncusOS version 202608102114
• Everything is running on Proxmox 8.4.1

Service configuration supplied to Operations Center at cluster creation:

linstor:
  enabled: true

Linstor controller has all the nodes Online:

# linstor node list -p
+--------------------------------------------------------------------+
| Node   | Platform | NodeType  | Addresses                | State   |
|====================================================================|
| node-1 | LINUX    | SATELLITE | 10.10.0.98:3366 (PLAIN)  | Online  |
| node-2 | LINUX    | SATELLITE | 10.10.0.44:3366 (PLAIN)  | Online  |
| node-3 | LINUX    | SATELLITE | 10.10.0.130:3366 (PLAIN) | Online  |
+--------------------------------------------------------------------+

Storage pools registered on the three storage-contributing members (backed by an IncusOS managed ZFS pool/volume):

linstor storage-pool create zfsthin node-1 nvme nvme/linstor
linstor storage-pool create zfsthin node-2 nvme nvme/linstor
linstor storage-pool create zfsthin node-3 nvme nvme/linstor

I can see the storage-pools available for each node:

# linstor storage-pool list -p
+-----------------------------------------------------------------------------------------------------------------------------------------------+
| StoragePool          | Node   | Driver   | PoolName     | FreeCapacity | TotalCapacity | CanSnapshots | State   | SharedName                  |
|===============================================================================================================================================|
| DfltDisklessStorPool | node-1 | DISKLESS |              |              |               | False        | Ok      | node-1;DfltDisklessStorPool |
| DfltDisklessStorPool | node-2 | DISKLESS |              |              |               | False        | Ok      | node-2;DfltDisklessStorPool |
| DfltDisklessStorPool | node-3 | DISKLESS |              |              |               | False        | Ok      | node-3;DfltDisklessStorPool |
| nvme                 | node-1 | ZFS_THIN | nvme/linstor |     6.69 GiB |      7.50 GiB | True         | Ok      | node-1;nvme                 |
| nvme                 | node-2 | ZFS_THIN | nvme/linstor |     6.69 GiB |      7.50 GiB | True         | Ok      | node-2;nvme                 |
| nvme                 | node-3 | ZFS_THIN | nvme/linstor |     6.69 GiB |      7.50 GiB | True         | Ok      | node-3;nvme                 |
+-----------------------------------------------------------------------------------------------------------------------------------------------+

The problem

Adding a 4th server via Operations Center:

Aug 11, 03:15:51 PM 6fded98a-4d96-4794-ada3-c6b25d422bbe operations-centerd[1015]: 2026-08-11T06:15:51Z ERR <= method JoinCluster returned an error err="Failed to wait for update operation during cluster join on \"node-1\" (https://10.10.0.175:8443): Failed to initialize member: Failed to initialize storage pools and networks: Failed to create storage pool \"linstor\": Could not determine DRBD module version: Module not loaded" request_id=be0c95c8-2632-42f1-a00a-4cabe708a04a
Aug 11, 03:15:51 PM 6fded98a-4d96-4794-ada3-c6b25d422bbe operations-centerd[1015]: 2026-08-11T06:15:51Z ERR <= method AddServers returned an error err="Failed to join cluster \"cluster\" for server \"nas\": Failed to wait for update operation during cluster join on \"node-1\" (https://10.10.0.175:8443): Failed to initialize member: Failed to initialize storage pools and networks: Failed to create storage pool \"linstor\": Could not determine DRBD module version: Module not loaded" request_id=be0c95c8-2632-42f1-a00a-4cabe708a04a
Aug 11, 03:15:51 PM 6fded98a-4d96-4794-ada3-c6b25d422bbe operations-centerd[1015]: 2026-08-11T06:15:51Z ERR Server error response method=POST request_uri=/1.0/provisioning/clusters/cluster/:add-servers ip=10.10.0.35:52956 status_code=500 response="Failed adding servers [nas] to cluster \"cluster\": Failed to join cluster \"cluster\" for server \"nas\": Failed to wait for update operation during cluster join on \"node-1\" (https://10.10.0.175:8443): Failed to initialize member: Failed to initialize storage pools and networks: Failed to create storage pool \"linstor\": Could not determine DRBD module version: Module not loaded" request_id=be0c95c8-2632-42f1-a00a-4cabe708a04a

Expected

A new member should join a cluster containing a linstor pool.

Does this happen because the OC applies service configuration after member initialization, so the DRBD module is not yet loaded when the pool is created?