Oversubscription in thin provisioning¶
OpenStack Block Storage enables you to choose a volume back end based on virtual capacities for thin provisioning using the oversubscription ratio.
A reference implementation is provided for the default LVM driver. The illustration below uses the LVM driver as an example.
Configure oversubscription settings¶
To support oversubscription in thin provisioning, a flag
max_over_subscription_ratio is introduced into
This is a float representation of the oversubscription ratio when thin
provisioning is involved. Default ratio is 20.0, meaning provisioned
capacity can be 20 times of the total physical capacity. A ratio of 10.5
means provisioned capacity can be 10.5 times of the total physical capacity.
A ratio of 1.0 means provisioned capacity cannot exceed the total physical
capacity. A ratio lower than 1.0 is ignored and the default value is used
This parameter also can be set as
using auto, Cinder will automatically calculate the
max_over_subscription_ratio based on the provisioned capacity and the used
space. This allows the creation of a larger number of volumes at the
beginning of the pool’s life, and start to restrict the creation as the free
space approaches to 0 or the reserved limit.
max_over_subscription_ratio can be configured for each back end when
multiple-storage back ends are enabled. It is provided as a reference
implementation and is used by the LVM driver. However, it is not a
requirement for a driver to use this option from
max_over_subscription_ratio is for configuring a back end. For a
driver that supports multiple pools per back end, it can report this
ratio for each pool. The LVM driver does not support multiple pools.
Setting this value to ‘auto’. The values calculated by Cinder can dynamically vary according to the pool’s provisioned capacity and consumed space.
reserved_percentage flag is used to prevent over provisioning.
This flag represents the percentage of the back-end capacity that is reserved.
There is a change on how
reserved_percentage is used. It was measured
against the free capacity in the past. Now it is measured against the total
Drivers can report the following capabilities for a back end or a pool:
thin_provisioning_support = True(or False) thick_provisioning_support = True(or False) provisioned_capacity_gb = PROVISIONED_CAPACITY max_over_subscription_ratio = MAX_RATIO
PROVISIONED_CAPACITY is the apparent allocated space indicating
how much capacity has been provisioned and
MAX_RATIO is the maximum
oversubscription ratio. For the LVM driver, it is
Two capabilities are added here to allow a back end or pool to claim support for thin provisioning, or thick provisioning, or both.
The LVM driver reports
thick_provisioning_support=False if the
lvm_type flag in
thin. Otherwise it reports
Volume type extra specs¶
If volume type is provided as part of the volume creation request, it can have the following extra specs defined:
'capabilities:thin_provisioning_support': '<is> True' or '<is> False' 'capabilities:thick_provisioning_support': '<is> True' or '<is> False'
capabilities scope key before
thick_provisioning_support is not required. So the following works too:
'thin_provisioning_support': '<is> True' or '<is> False' 'thick_provisioning_support': '<is> True' or '<is> False'
The above extra specs are used by the scheduler to find a back end that supports thin provisioning, thick provisioning, or both to match the needs of a specific volume type.
Volume replication extra specs¶
OpenStack Block Storage has the ability to create volume replicas. Administrators can define a storage policy that includes replication by adjusting the cinder volume driver. Volume replication for OpenStack Block Storage helps safeguard OpenStack environments from data loss during disaster recovery.
To enable replication when creating volume types, configure the cinder
Each volume created with the replication capability set to
generates a copy of the volume on a storage back end.
One use case for replication involves an OpenStack cloud environment installed across two data centers located nearby each other. The distance between the two data centers in this use case is the length of a city.
At each data center, a cinder host supports the Block Storage service. Both data centers include storage back ends.
Depending on the storage requirements, there can be one or two cinder
hosts. The administrator accesses the
/etc/cinder/cinder.conf configuration file and sets
If one data center experiences a service failure, administrators can redeploy the VM. The VM will run using a replicated, backed up volume on a host in the second data center.
In the capacity filter,
max_over_subscription_ratio is used when
choosing a back end if
thin_provisioning_support is True and
max_over_subscription_ratio is greater than 1.0.
In the capacity weigher, virtual free capacity is used for ranking if
thin_provisioning_support is True. Otherwise, real free capacity
will be used as before.