openstack-manuals/doc/admin-guide-cloud/source/blockstorage_multi_backend.rst
willy lin 83e642a05a Change default of option volume_driver to LVMVolumeDriver
message:
"LVMISCSIDriver" change to "LVMVolumeDriver"

Change-Id: I8f1a5bc63ba108f804e39289e2a201fbfb517b7e
Closes-Bug: #1483081
2015-08-27 19:16:28 +02:00

6.1 KiB

Configure multiple-storage back ends

When you configure multiple-storage back ends, you can create several back-end storage solutions that serve the same OpenStack Compute configuration and one cinder-volume is launched for each back-end storage or back-end storage pool.

In a multiple-storage back-end configuration, each back end has a name (volume_backend_name). Several back ends can have the same name. In that case, the scheduler properly decides which back end the volume has to be created in.

The name of the back end is declared as an extra-specification of a volume type (such as, volume_backend_name=LVM). When a volume is created, the scheduler chooses an appropriate back end to handle the request, according to the volume type specified by the user.

Enable multiple-storage back ends

To enable a multiple-storage back ends, you must set the enabled_backends flag in the cinder.conf file. This flag defines the names (separated by a comma) of the configuration groups for the different back ends: one name is associated to one configuration group for a back end (such as, [lvmdriver-1]).

Note

The configuration group name is not related to the volume_backend_name.

Note

After setting the enabled_backends flag on an existing cinder service, and restarting the Block Storage services, the original host service is replaced with a new host service. The new service appears with a name like host@backend. Use:

$ cinder-manage volume update_host --currentname CURRENTNAME --newname CURRENTNAME@BACKEND

to convert current block devices to the new hostname.

The options for a configuration group must be defined in the group (or default options are used). All the standard Block Storage configuration options (volume_group, volume_driver, and so on) might be used in a configuration group. Configuration values in the [DEFAULT] configuration group are not used.

These examples show three back ends:

enabled_backends=lvmdriver-1,lvmdriver-2,lvmdriver-3
[lvmdriver-1]
volume_group=cinder-volumes-1
volume_driver=cinder.volume.drivers.lvm.LVMVolumeDriver
volume_backend_name=LVM
[lvmdriver-2]
volume_group=cinder-volumes-2
volume_driver=cinder.volume.drivers.lvm.LVMVolumeDriver
volume_backend_name=LVM
[lvmdriver-3]
volume_group=cinder-volumes-3
volume_driver=cinder.volume.drivers.lvm.LVMVolumeDriver
volume_backend_name=LVM_b

In this configuration, lvmdriver-1 and lvmdriver-2 have the same volume_backend_name. If a volume creation requests the LVM back end name, the scheduler uses the capacity filter scheduler to choose the most suitable driver, which is either lvmdriver-1 or lvmdriver-2. The capacity filter scheduler is enabled by default. The next section provides more information. In addition, this example presents a lvmdriver-3 back end.

Note

For Fiber Channel drivers that support multipath, the configuration group requires the use_multipath_for_image_xfer=true option. In the example below, you can see details for HP 3PAR and EMC Fiber Channel drivers.

[3par]
use_multipath_for_image_xfer = true
volume_driver = cinder.volume.drivers.san.hp.hp_3par_fc.HP3PARFCDriver
volume_backend_name = 3parfc

[emc]
use_multipath_for_image_xfer = true
volume_driver = cinder.volume.drivers.emc.emc_smis_fc.EMCSMISFCDriver
volume_backend_name = emcfc

Configure Block Storage scheduler multi back end

You must enable the filter_scheduler option to use multiple-storage back ends. The filter scheduler:

  1. Filters the available back ends. By default, AvailabilityZoneFilter, CapacityFilter and CapabilitiesFilter are enabled.
  2. Weights the previously filtered back ends. By default, the CapacityWeigher option is enabled. When this option is enabled, the filter scheduler assigns the highest weight to back ends with the most available capacity.

The scheduler uses filters and weights to pick the best back end to handle the request. The scheduler uses volume types to explicitly create volumes on specific back ends.

Volume type

Before using it, a volume type has to be declared to Block Storage. This can be done by the following command:

$ cinder --os-username admin --os-tenant-name admin type-create lvm

Then, an extra-specification has to be created to link the volume type to a back end name. Run this command:

$ cinder --os-username admin --os-tenant-name admin type-key lvm set \
  volume_backend_name=LVM_iSCSI

This example creates a lvm volume type with volume_backend_name=LVM_iSCSI as extra-specifications.

Create another volume type:

$ cinder --os-username admin --os-tenant-name admin type-create lvm_gold

$ cinder --os-username admin --os-tenant-name admin type-key lvm_gold set \
  volume_backend_name=LVM_iSCSI_b

This second volume type is named lvm_gold and has LVM_iSCSI_b as back end name.

Note

To list the extra-specifications, use this command:

$ cinder --os-username admin --os-tenant-name admin extra-specs-list

Note

If a volume type points to a volume_backend_name that does not exist in the Block Storage configuration, the filter_scheduler returns an error that it cannot find a valid host with the suitable back end.

Usage

When you create a volume, you must specify the volume type. The extra-specifications of the volume type are used to determine which back end has to be used.

$ cinder create --volume_type lvm --display_name test_multi_backend 1

Considering the cinder.conf described previously, the scheduler creates this volume on lvmdriver-1 or lvmdriver-2.

$ cinder create --volume_type lvm_gold --display_name test_multi_backend 1

This second volume is created on lvmdriver-3.