OpenStack Nova Compute & Scheduling Errors
Everything that stops an OpenStack instance from booting or moving: the scheduler finds No valid host, Placement returns no allocation candidates, an instance fails to spawn or gets stuck, a live migration aborts, or a NUMA/PCI request can't fit. Each guide opens with the direct cause and the `openstack`, `nova-manage`, and placement commands to confirm it. Start with your symptom below, or paste the nova-scheduler log into the assistant.
Fix your error now
Paste the error or logs and let the Incident Assistant diagnose it, or run a guided triage prompt with AI — no copy-paste.
Every guide in this cluster
19 guides-
'No valid host was found' in OpenStack Nova — Causes & Fix
Fix OpenStack Nova's 'No valid host was found' scheduling error: usually exhausted Placement capacity or a strict filter. Diagnose flavors, ratios, and fix it.
-
'Instance failed to spawn' Nova Stuck in BUILD/spawning
Fix Nova 'Instance failed to spawn' and instances stuck in BUILD/spawning: diagnose libvirt/qemu errors, disk space, VIF plug timeouts, SELinux, and CPU flags.
-
'No allocation candidates returned' Placement API
Fix the Placement 'No allocation candidates returned' error: reconcile inventory vs usage, traits, host/placement aggregate mismatch, and heal stale allocations.
-
Nova Migration Pre-Check Failed (Host, CPU, or Disk)
Fix Nova live/cold migration 'MigrationPreCheckError' failures: diagnose incompatible CPU flags, insufficient host disk, no valid destination, and shared-storage or flavor mismatches.
-
'Build of instance aborted' Nova Spawn Failure
Fix the Nova 'Build of instance aborted' error: distinguish aborted vs rescheduled failures and trace network, volume/BDM, and image causes in compute logs.
-
'Filtering removed all hosts' Nova Filter Chain
Fix Nova 'Filtering removed all hosts' / 'Filter returned 0 hosts': read scheduler debug logs, find which filter zeroed the list, and inspect per-filter counts.
-
'InstanceNotFound' nova-compute manager failure
Nova logging InstanceNotFound during periodic tasks or deletes? Diagnose orphaned database rows, stale local instances, and compute-manager drift step by step.
-
'No space left on device' — Fix Nova Compute Disk Exhaustion
Fix Nova 'No space left on device' errors on compute hosts: diagnose full instances/image-cache partitions, disk_allocation_ratio overcommit, orphaned disks, and log growth in Kolla-Ansible OpenStack.
-
'Placement 409 Conflict' allocation update failure
Nova logging Placement 409 Conflict on inventory or allocation updates? Diagnose generation conflicts, stale resource providers, and concurrent writes step by step.
-
'task_state deleting' — Instance Stuck Deleting in Nova
Fix an OpenStack instance hung in task_state 'deleting': diagnose stuck volume detach, Neutron port cleanup, RabbitMQ RPC failures, and safely use nova reset-state and delete --force.
-
Nova 'Live Migration failure' / Migration Failed
Fix Nova live migration failures in OpenStack: resolve CPU model mismatches, missing shared storage, firewall-blocked libvirt ports, timeouts, and NUMA pinning.
-
nova-compute Service State 'down' / Hypervisor Down
Fix the nova-compute 'down' hypervisor state in OpenStack: diagnose dead services, RabbitMQ drops, clock skew, stale placement, and force-down for evacuation.
-
Placement 'Resource provider with uuid not found' / 404
Fix the Placement 'resource provider not found' 404 in OpenStack: repair stale RPs after host renames, mismatched hostnames, orphaned allocations, and re-added nodes.
-
'Cannot fit instance NUMA topology' NUMA/CPU Pinning Scheduling Failure
Fix Nova's 'Requested instance NUMA topology cannot fit the given host NUMA topology' scheduling failure: diagnose CPU pinning, hugepages, and NUMA host capacity.
-
'Flavor ... could not be found' Launching an Instance
Fix OpenStack Nova 'Flavor could not be found' (HTTP 404) when booting a server: diagnose deleted flavors, wrong IDs vs names, and missing project flavor access.
-
'PCI device not found' SR-IOV / PCI Passthrough Scheduling Failure
Fix Nova's 'PCI device not found' / no matching PCI alias error for SR-IOV and passthrough: verify device_spec, alias, IOMMU, and PciPassthroughFilter placement.
-
Nova Instance Stuck in ERROR State With a Fault Trace
Fix a Nova instance in ERROR with a fault in 'openstack server show': diagnose NoValidHost, libvirt spawn failures, and image/quota faults, then rebuild or reset-state.
-
nova-compute 'Error updating resources' — fix resource tracker claim failures
Fix nova-compute resource tracker exceptions and 'Error updating resources for node': diagnose Placement sync, stale allocations, and ComputeResourcesUnavailable claim failures on the compute host.
-
Placement 'inventory in use' 409 Conflict Updating Resource Provider
Fix Placement 409 'update conflict: Inventory for ... in use' when editing resource provider inventory: diagnose existing allocations, orphaned consumers, and safely adjust capacity.
OpenStack AI prompts
Copy-paste, production-safe prompts for this stack.
OpenStack command center
Top errors, validators, and runbooks for the whole stack.
Frequently asked questions
What does "No valid host was found" actually mean?
nova-scheduler logs for which filter removed the hosts, and openstack resource provider inventory list. See No valid host was found.Why does Placement return "no allocation candidates"?
openstack allocation candidate list to see what Placement thinks is available.