Hardware Overview
The Hardware section shows the physical state of the appliance.
Cluster Map
Hardware → Cluster Map shows one card per node with its drive bays and network link status together in a single view. Drive bays are laid out as vertical slots resembling a server front panel, and shared or system drives are marked so you can see how each drive is attributed.
Click a node for a detail panel with CPU, memory, drive, and link vitals. Click a device for its detail dialog, which includes a Temperature tab.
Nodes
Hardware → Nodes lists each node with its phase, software version, and reachability status.
| Reachability | Meaning |
|---|---|
| Synced | Normal |
| Recovering | Storage is rejoining on the fabric; self-clears |
| Unreachable | Storage cannot be reached on the fabric — check the link |
A node that drops out of the cluster network rejoins on its own; a manual service restart is not needed.
When the cluster is not Ready, a Diagnostics panel appears here showing the cluster message, failed conditions, and per-node phase and heartbeat detail.
Drives
Hardware → Drives lists every NVMe device with its vendor, model, capacity, health, and whether it is enrolled in the storage cluster.
| Column / chip | Meaning |
|---|---|
| Drive Status | Healthy, Failed, Unknown, or Empty — the help icon on the column explains each |
| In Cluster | Whether the drive is enrolled in the storage cluster |
| System Disk | The drive backs the operating system — detected even when the mount sits on LVM, LUKS, or MD-RAID. Never usable as cluster storage. |
Each node's drive inventory refreshes itself regularly, so the list does not show drives that are gone or offer stale drives for import.
A drive fault does not take down the whole node: only the faulty drive is removed from service, data rebuild starts, and the rest of the node keeps serving. To replace a drive, see Replacing a Drive.
A non-zero lifetime "media and data integrity errors" count does not by itself mark a drive failed. The raw counter is still published for graphs and alerts.
Temperature Trends
Device detail dialogs on the Cluster Map and Drives pages have a Temperature tab with per-device graphs over the last 1, 6, or 24 hours, including warning and critical reference lines, so you can spot an overheating drive or network card before it fails.
Network
Hardware → Network shows the cluster's network interfaces and subnets, including a Fabric Protocol chip (RDMA or TCP) so the transport is visible at a glance. Interfaces that are link-down or not bound to any subnet are listed too, with separate Up / Down / Unknown / Unused status chips.
See Network Concepts.