Compute Hosts

Overview

Compute hosts are the physical bare-metal nodes where virtual machine instances run. Each host runs a Compute Agent that communicates with the local hypervisor and reports resource availability back to the control plane. Administrators manage compute hosts to maintain cluster health, drain nodes before maintenance, and monitor resource allocation across the fleet.

Administrator Access Required — This operation requires the admin role. Contact your Xloud administrator if you do not have sufficient permissions.

Prerequisites

Admin credentials sourced from admin-openrc.sh
openstack CLI installed and configured
Xloud Compute services running and healthy

Host Metrics

The following metrics are available for every registered compute host.

Metric	Description
Status	`enabled` / `disabled` — whether the scheduler places new instances on this host
State	`up` / `down` — connectivity between the Compute Agent and the API tier
vCPUs Used / Total	Allocated vCPU threads vs. total physical threads on the host
Memory Used / Total	Allocated RAM (MB) vs. installed RAM on the host
Disk Used / Total	Storage consumed (GB) vs. available local storage
Running Instances	Count of active instances currently on this host

Status and State are independent. A host can be enabled (accepts new placements) but down (agent is unreachable). Always verify both fields when diagnosing capacity issues.

List and Inspect Hosts

Dashboard
CLI
XDeploy

Navigate to host management

Log in to the Xloud Dashboard (https://connect.<your-domain>) and navigate to Admin → Compute → Hypervisors.The hypervisor list displays all registered compute nodes with real-time capacity metrics. The summary row at the top shows aggregate cluster utilization.

Review host metrics

Each row shows the host’s status, state, vCPU allocation, memory allocation, disk usage, and running instance count.Click any host name to open its detail panel, which shows a breakdown of all running instances and their resource consumption.

Authenticate

Source your credentials file to authenticate with the Xloud platform:

Load credentials

source admin-openrc.sh

Download the OpenRC file from Xloud Dashboard → Project → API Access → Download OpenStack RC File.

List all compute hosts

List hypervisors

openstack hypervisor list --long

Filter by status to identify degraded hosts:

Filter disabled hosts

openstack hypervisor list --long | grep -i disabled

Show host details

Inspect a specific host

openstack hypervisor show <hostname-or-id>

The output includes current vCPU, memory, disk usage, and the full list of running instances on that node.

Verify that state is up and status is enabled for all production hosts.

Enable and Disable Hosts

Disabling a host prevents the scheduler from placing new instances on it. Existing instances continue running uninterrupted. Use this procedure before performing host maintenance.

Dashboard
CLI

Locate the host

Navigate to Admin → Compute → Hypervisors and locate the host you want to disable.

Disable the host

Click the host name → Disable Host.The scheduler immediately stops placing new instances on this host. The status column changes to disabled.

Disabling a host does not migrate existing instances. Drain the host by live-migrating all running instances before performing maintenance that requires a reboot. See Live Migration for instructions.

Re-enable after maintenance

When maintenance is complete, navigate back to the host detail and click Enable Host.

The host status returns to enabled and the scheduler resumes placing instances on it.

Disable a host

Disable host for maintenance

openstack compute service set \
  --disable \
  --disable-reason "Scheduled maintenance" \
  <hostname> nova-compute

Verify the status change

Confirm host is disabled

openstack hypervisor list --long | grep <hostname>

The Status column shows disabled.

Re-enable after maintenance

Re-enable host after maintenance

openstack compute service set --enable <hostname> nova-compute

Confirm re-enabled

openstack hypervisor list --long | grep <hostname>

The Status column returns to enabled.

Resource Allocation Overview

Xloud Compute supports over-commitment — the scheduler allows the sum of allocated resources to exceed the host’s physical capacity based on configurable ratios.

Resource	Default Over-Commit Ratio	Example (32-core host)
vCPU	16:1	Up to 512 vCPUs schedulable
RAM	1.5:1	128 GB RAM → up to 192 GB schedulable
Disk	1.0:1	No over-commit by default

In production environments, tune over-commit ratios based on actual workload profiles. High over-commitment on memory-intensive workloads can cause host-level out-of-memory conditions. Configure ratios through XDeploy under Compute → Advanced Settings → Resource Allocation.

Next Steps

Live Migration

Drain hosts by live-migrating running instances before maintenance.

Scheduling

Configure filters and weighers that control host selection.

Admin Guide

Return to the Compute Administration Guide index.

Core Services

Other Services

Compute Hosts

Overview

Host Metrics

List and Inspect Hosts

Navigate to host management

Review host metrics

Authenticate

List all compute hosts

Show host details

Enable and Disable Hosts

Locate the host

Disable the host

Re-enable after maintenance

Disable a host

Verify the status change

Re-enable after maintenance

Resource Allocation Overview

Next Steps

Live Migration

Scheduling

Admin Guide

Core Services

Other Services

​Overview

​Host Metrics

​List and Inspect Hosts

Navigate to host management

Review host metrics

Authenticate

List all compute hosts

Show host details

​Enable and Disable Hosts

Locate the host

Disable the host

Re-enable after maintenance

Disable a host

Verify the status change

Re-enable after maintenance

​Resource Allocation Overview

​Next Steps

Live Migration

Scheduling

Admin Guide

Overview

Host Metrics

List and Inspect Hosts

Enable and Disable Hosts

Resource Allocation Overview

Next Steps