OpenStack Stacking vs Spreading

Compute (Nova)
2026-04-17T07:12:20+00:00
Configure OpenStack scheduler weights and overcommit ratios to choose between stacking VMs for density or spreading them for isolation.

OpenStack Stacking vs Spreading

Set ram_weight_multiplier to -1.0 for stacking or 1.0 for spreading in /etc/nova/nova.conf.

Configuration

Edit /etc/nova/nova.conf on the controller node.

Stacking (High Density)

Pack instances onto fewer hosts.

[DEFAULT]
scheduler_weight_classes = nova.scheduler.weights.ram.RAMWeigher

[filter_scheduler]
enabled_filters = AvailabilityZoneFilter,ComputeFilter,ComputeCapabilitiesFilter,ImagePropertiesFilter,ServerGroupAntiAffinityFilter,ServerGroupAffinityFilter,NUMATopologyFilter
ram_weight_multiplier = -1.0

Spreading (High Isolation)

Distribute instances evenly across hosts.

[DEFAULT]
scheduler_weight_classes = nova.scheduler.weights.ram.RAMWeigher

[filter_scheduler]
enabled_filters = AvailabilityZoneFilter,ComputeFilter,ComputeCapabilitiesFilter,ImagePropertiesFilter,ServerGroupAntiAffinityFilter,ServerGroupAffinityFilter,NUMATopologyFilter
ram_weight_multiplier = 1.0

Multi-Resource Weighing

Combine RAM, CPU, and Disk weighers.

[DEFAULT]
scheduler_weight_classes = nova.scheduler.weights.ram.RAMWeigher,nova.scheduler.weights.cpu.CPUWeigher,nova.scheduler.weights.disk.DiskWeigher

[filter_scheduler]
ram_weight_multiplier = 1.0
cpu_weight_multiplier = 1.0
disk_weight_multiplier = 1.0

Overcommit Ratios

Adjust virtual-to-physical ratios in /etc/nova/nova.conf.

[DEFAULT]
cpu_allocation_ratio = 4.0
ram_allocation_ratio = 1.5
disk_allocation_ratio = 1.0

Deployment Scenarios

Dev/Test (Stacking + High Overcommit)

Maximize density for non-critical workloads.

[filter_scheduler]
ram_weight_multiplier = -1.0

[DEFAULT]
cpu_allocation_ratio = 16.0
ram_allocation_ratio = 2.0

Production (Spreading + Low Overcommit)

Prioritize performance isolation.

[filter_scheduler]
ram_weight_multiplier = 1.0

[DEFAULT]
cpu_allocation_ratio = 2.0
ram_allocation_ratio = 1.0

Mixed Strategy via Host Aggregates

Apply different overcommit policies to specific host groups.

openstack aggregate create --availability-zone dev-zone dev-aggregate
openstack aggregate create --availability-zone prod-zone prod-aggregate
openstack aggregate add host dev-aggregate compute-dev-01
openstack aggregate add host prod-aggregate compute-prod-01
openstack aggregate add host prod-aggregate compute-prod-02
openstack aggregate set --property cpu_allocation_ratio=16.0 dev-aggregate
openstack aggregate set --property cpu_allocation_ratio=2.0 prod-aggregate

Verification

Check instance placement:

openstack server list --long -c Name -c Host -c Status

Review hypervisor resource usage:

openstack hypervisor list --long
openstack hypervisor show compute-01 -c vcpus_used -c memory_mb_used -c running_vms

Troubleshooting

  • All VMs on one host: Set ram_weight_multiplier = 1.0.
  • No valid host found: Check enabled_filters, resource availability, and overcommit limits.
  • Uneven distribution: Verify scheduler_weight_classes and multiplier signs.
  • OOM kills: Reduce ram_allocation_ratio or increase physical RAM.