Hosts

Path: Left sidebar > Infrastructure > Hosts

Host Inventory Overview

The Hosts page lists every node that participates in the Ceph cluster, along with its hardware profile and the daemons it runs.

What This Screenshot Shows: Hosts Dashboard (UI Reference; Values Depend On Your Environment).

Hosts List - Column Reference

Column	What It Shows
Hostname	The name of the node as registered in the cluster (FQDN or short name)
Services	List of daemon types running on this host (for example osd, mgr, mds, rgw)
Labels	Ceph placement labels assigned to this host, used by the orchestrator to decide which services to place here
Address	Primary network IP address of this host
Vendor / Model	Hardware vendor and server model from BMC or DMI data. -- means not reported.
CPUs	Number of physical CPU sockets
Cores	Total physical CPU cores
Total Memory	Total installed RAM
Raw Capacity	Sum of disk capacity on this host
HDDs	Number of spinning hard drives
Flash	Number of SSDs and NVMe drives
NICs	Number of network interface cards
Status	Host reachability and maintenance state.
Actions	Edit labels, enter maintenance, remove host

Tip

Services badges show daemon type and count per host. Example: osd:4 means 4 OSD daemons are running on that host.

Tip

Service badge colors help quick scanning. Example mappings include: OSD (orange), RGW (blue), MDS (teal), MGR (green), and crash (pink).

Tip

Labels tell the Ceph orchestrator which hosts are eligible for each service type. A host without the correct label will not receive that service.

What You See

Tab	Description
Daemons	Lists all Ceph daemons running on the host with type, ID, version, and current status.
Devices	Shows all block devices visible to Ceph on this host, including device path, type, available status, and OSD ID if in use.
Device Health (SMART)	SMART telemetry for every disk on the host. Use wear level and reallocated sectors to detect early disk failure risk.
Performance	Per-host CPU, memory, and network utilization metrics from Prometheus node_exporter, shown as utilization panels/cards in this UI.

How To Filter Hosts By Service

The Services filter at the top-left of the Hosts list lets you show only hosts running a selected daemon type.

Purpose:

• To quickly identify which hosts run a specific service.

• To verify service placement after a deployment change.

When to Use:

• Before changing labels or maintenance state.

• During service troubleshooting and placement checks.

Steps:

1. Open Infrastructure > Hosts.

2. Click the Services filter (funnel icon).

3. Select one or more service type checkboxes (for example mon, mgr, osd, mds).

4. Review the filtered host list.

5. Clear all checkboxes to return to full list.

Expected Outcome:

• You isolate host rows to only the service types you need to inspect.

What This Screenshot Shows: Hosts Service Filter (UI Reference; Values Depend On Your Environment).

How To Add A Host

Path: Infrastructure > Hosts > + Add Host

Prerequisite:

Before adding a host, your administrator must configure the new host with the cephadm SSH key. The cluster uses SSH to manage hosts.

Purpose:

• To register a new physical node so it can run Ceph daemons.

• To expand cluster capacity and future placement options.

When to Use:

• During cluster expansion.

• Before deploying services on a newly provisioned host.

Steps:

1. Open Infrastructure > Hosts.

2. Click + Add Host.

3. Enter Hostname.

4. Enter Address.

5. Add Labels if required.

6. Leave Maintenance Mode unchecked for normal registration.

7. Click Add Host.

Expected Outcome:

• Host appears in list within seconds.

• Orchestrator connects via SSH and discovers hardware profile.

• CPU, memory, capacity, and device counts populate.

• Host becomes eligible for service deployment based on labels.

What This Screenshot Shows: Add Host Panel (UI Reference; Values Depend On Your Environment).

Field	Value / Options	Description
Hostname *	Text input	Required. Must match the node hostname.
Address *	Text input	Required. IP address used by the cluster.
Labels	Text tags	Optional. Controls service placement eligibility.
Maintenance Mode	Checkbox	Optional. Puts host in maintenance mode immediately.

Common Labels Reference:

Label	Purpose
osd	Host is eligible for OSD placement
mon	Host is eligible for Monitor placement
mgr	Host is eligible for Manager placement
rgw	Host is eligible for Gateway placement
mds	Host is eligible for MDS placement
_no_schedule	Prevents placement of new daemons on this host
_admin	Host receives admin keyring

Host Row Actions - ... Menu

Click the row Actions menu on the right of a host row.

What This Screenshot Shows: Host Row Actions Menu (UI Reference; Values Depend On Your Environment).

Edit Labels

Purpose:

• Opens label editor for this host.

Steps:

1. Click ... on host row.

2. Click Edit Labels.

3. Add/remove labels.

4. Click Save.

Expected Outcome:

• Orchestrator immediately uses updated labels for placement decisions.

• Adding a label can make the host eligible for new service deployment.

• Removing a label can trigger daemon migration off this host.

Warning

Removing critical labels from essential hosts can destabilize placement. Review impact before removing labels.

Maintenance Mode

Purpose:

• Places host into maintenance mode for planned hardware work.

Steps:

1. Click ... on host row.

2. Click Maintenance Mode.

3. Confirm dialog.

What Happens During Maintenance:

• OSDs on this host are marked out and data rebalances.

• Non-essential daemons are stopped.

• Host state reflects maintenance mode.

Expected Outcome:

• Host is safe for planned physical work after rebalancing completes.

Important

Do not power off the host until rebalancing is complete and the Dashboard returns to active+clean PG state.

Warning

Do not put multiple hosts into maintenance mode simultaneously unless replication safety on remaining hosts is verified.

Remove Host

Purpose:

• Deregisters a host from the cluster after daemons are stopped.

Steps:

1. Verify all non-essential daemons are stopped on the host.

2. Verify all OSDs on that host are removed or down+out.

3. Click ... on host row.

4. Click Remove Host.

5. Confirm removal dialog.

Expected Outcome:

• Host is removed from Ceph host inventory and no longer receives deployments.

Warning

Remove Host requires all OSDs on that host to be removed or down+out. Do not remove hosts still serving active data placement.

Host Detail Tabs

Click chevron > to expand a host row.

Daemons Tab

Lists all Ceph daemons currently running on this host.

Column	What It Shows
Daemon name	Full daemon identifier
Version	Ceph version running for that daemon instance
Status	running means healthy
Last Refreshed	Time since last status update
CPU Usage	Current CPU usage percent
Memory Usage	Current daemon memory usage
Daemon Events	Recent deployment/restart/error events. No data available means no recent events.

Steps:

1. Open Daemons tab.

2. Verify all daemons show running.

3. Inspect recent daemon events for non-running instances.

What This Screenshot Shows: Host Expanded Row - Daemons Tab (UI Reference; Values Depend On Your Environment).

Devices Tab

Shows all block devices visible on this host.

Column	What It Shows
Device path	Kernel path such as /dev/sda or /dev/nvme1n1
Type	Device media type
Available	Yes means ready for new OSD deployment
Vendor / Model	Drive vendor and model identifiers
Size	Raw disk capacity
OSDs	OSD ID if device is already in use

Steps:

1. Open Devices tab.

2. Confirm media type detection is correct.

3. Check Available for OSD-ready disks.

What This Screenshot Shows: Host Expanded Row - Devices Tab (UI Reference; Values Depend On Your Environment).

Device Health Tab

Shows SMART telemetry data for each disk on this host.

Steps:

1. Open Device Health tab.

2. Select disk tab at top.

3. Check SMART overall-health result.

4. Review temperature and power-on hours.

5. Open Device Information for full SMART summary fields.

6. Open SMART for raw SMART attribute data when deeper diagnostics are needed.

Key Values To Monitor:

Value	What To Check
SMART overall-health	Must show passed
Health	PASSED indicates healthy disk
Temp	Temperature within drive operating range
Power-on hrs	Use with wear indicators for lifecycle planning
Wear Level	Check SSD/NVMe wear trend. Rapidly degrading wear indicates disk aging risk.
Reallocated Sectors	Any growth trend can indicate media degradation and early failure risk.

Warning

Disks reporting SMART failure should be replaced before full failure.

What This Screenshot Shows: Host Expanded Row - Device Health Tab (UI Reference; Values Depend On Your Environment).

Performance Tab

Shows per-host CPU, memory, and network utilization metrics sourced from Prometheus node_exporter. In this UI, those metrics are displayed as utilization panels/cards for quick host-level validation.

Card	What It Shows
CPU	Processor model and socket/core count
Memory	Total installed RAM
Services	Total number of daemon services on this host
Address	Host IP address
Storage	Disk count and capacity breakdown
NICs	Number of network interfaces
Ceph Version	Ceph version reported for this host (-- means not yet reported)
Source	Registration source (for example orchestrator-managed host)

What This Screenshot Shows: Host Expanded Row - Performance Tab (UI Reference; Values Depend On Your Environment).

Troubleshooting - Hosts

Problem You See	Most Likely Cause	What To Do
Host Status shows Offline	Host unreachable	Verify host power/network and check daemon health
Host added but hardware fields show dashes	Discovery still in progress	Wait a few minutes and refresh
Add Host fails	cephadm SSH key not configured	Configure cephadm SSH key on target host first
Services filter shows fewer hosts than expected	Some hosts do not run selected service	Clear/change filter to verify all hosts
Maintenance mode takes long time	Large rebalance in progress	Wait until PGs return to active+clean
Remove Host fails	One or more OSDs are still active on the host (not removed or not down+out)	Move host to maintenance, verify OSDs are removed or down+out, then retry remove

Note

If any issue persists, raise a support ticket via Monitoring > Alerts or Karios Support.