What should I do with HA disk slot inconsistencies after disk roaming?
Applicable Products
- QuTS hero h5.3.0 or later
- High Availability Manager
Background
QNAP NAS devices support disk roaming, which allows users to move disks to different slots while the system is powered off. After restarting, the system can still detect and mount the storage pool that the moved disks belong to, even though their locations have changed.
However, if your NAS devices are in a high-availability (HA) cluster, performing disk roaming only on one node device may cause High Availability Manager to detect disk slot inconsistencies between the two nodes. These inconsistencies may affect cluster health and stability.
Scenarios
The effect of disk roaming on the HA cluster depends on where you moved your disks on one node relative to the other node.
There are two major disk roaming scenarios. Let’s assume you powered off the passive node and moved two disks that belong to the same storage pool to different disk slots:
| Scenario | Active Node | Passive Node | Description |
|---|---|---|---|
| Swapped slots | Slot 1: Disk A (400 GB) Slot 2: Disk B (200 GB) | Slot 1: Disk B (200 GB) Slot 2: Disk A (400 GB) | The two disks on the passive node swapped slots compared to the corresponding disks on the active node, but the same two slots are still occupied. |
| Different slots | Slot 1: Disk A (400 GB) Slot 2: Disk B (200 GB) | Slot 3: Disk A (400 GB) Slot 4: Disk B (200 GB) | The two disks on the passive node were moved to different slots than the corresponding disks on the active node. |
The effects of each scenario on the HA status are as follows:
| Scenario | HA Status | Details |
|---|---|---|
| Swapped slots | Good | The storage pool and data synchronization still work as expected. However, future actions involving the disk slots may be affected (see important note in next section). |
| Different slots | Warning | Disk slot mismatch is detected. The storage pool and data synchronization still work as expected. However, cluster health is affected, and future actions involving the disk slots may be affected (see important note in next section). |
Solution
In both scenarios above, the HA cluster will continue to operate and sync data between the two nodes normally.
However, we highly recommend that the corresponding disks are always installed in matching disk slots between the two nodes. This ensures the long-term health and stability of the HA cluster.
You can resolve the disk slot inconsistency by doing one of the following:
- Perform a switchover, shut down the passive node, and then move the corresponding disks on the original active node to the same slots as those on the original passive node.
- Move the disks on the passive node back to their original disk slots.
If you leave the disk slot inconsistency in place, you may fail to perform future storage actions involving those disks and disk slots.
For example, if you remove the storage pool and try to create a new one using the same disks while a scenario of swapped slots or different slots is still present between the two nodes, High Availability Manager will recheck the hardware configuration of both nodes and detect the inconsistency, preventing the system from creating the new storage pool.
Further Reading
How do I replace a hard drive with a larger capacity in an HA cluster?
