Todo
rework after iSCSI merge (see ‘Old Docs’) (todd or vish)
Volume manager manages creating, attaching, detaching, and persistent storage.
Persistent storage volumes keep their state independent of instances. You can attach to an instance, terminate the instance, spawn a new instance (even one from a different image) and re-attach the volume with the same data intact.
Related Flags
volume_topic: | What rpc topic to listen to (default: cinder-volume). |
---|---|
volume_manager: | The module name of a class derived from manager.Manager (default: cinder.volume.manager.Manager). |
volume_driver: | Used by Manager. Defaults to cinder.volume.drivers.lvm.LVMISCSIDriver. |
volume_group: | Name of the group that will contain exported volumes (default: cinder-volumes) |
num_shell_tries: | |
Number of times to attempt to run commands (default: 3) |
Bases: cinder.manager.SchedulerDependentManager
Manages attachable block storage devices.
Updates db to show volume is attached.
Uploads the specified volume to Glance.
image_meta is a dictionary containing the following keys: ‘id’, ‘container_format’, ‘disk_format’
Creates the cgsnapshot.
Creates the consistency group.
Creates the consistency group from source.
Currently the source can only be a cgsnapshot.
Creates and exports the snapshot.
Creates the volume.
Deletes cgsnapshot.
Deletes consistency group and the volumes in the group.
Perform any required initialization.
Prepare volume for connection from host represented by connector.
This method calls the driver initialize_connection and returns it to the caller. The connector parameter is a dictionary with information about the host that will connect to the volume in the following format:
{
'ip': ip,
'initiator': initiator,
}
ip: the ip address of the connecting machine
initiator: the iscsi initiator name of the connecting machine. This can be None if the connecting machine does not support iscsi connections.
driver is responsible for doing any necessary security setup and returning a connection_info dictionary in the following format:
{
'driver_volume_type': driver_volume_type,
'data': data,
}
Migrate the volume to the specified host (called on source host).
Promote volume replica secondary to be the primary volume.
Collect driver status and then publish.
Re-enable replication of secondary volume with primary volumes.
Cleanup connection from host represented by connector.
The format of connector is the same as for initialize_connection.
Updates consistency group.
Update consistency group by adding volumes to the group, or removing volumes from the group.
Finalize migration process on backend device.
Lock decorator for volume detach operations.
Takes a named lock prior to executing the detach call. The lock is named with the operation executed and the id of the volume. This lock can then be used by other operations to avoid operation conflicts on shared volumes.
This locking mechanism is only for detach calls. We can’t use the locked_volume_operation, because detach requires an additional attachment_id in the parameter list.
Lock decorator for snapshot operations.
Takes a named lock prior to executing the operation. The lock is named with the operation executed and the id of the snapshot. This lock can then be used by other operations to avoid operation conflicts on shared snapshots.
Example use:
If a snapshot operation uses this decorator, it will block until the named lock is free. This is used to protect concurrent operations on the same snapshot e.g. delete SnapA while create volume VolA from SnapA is in progress.
Lock decorator for volume operations.
Takes a named lock prior to executing the operation. The lock is named with the operation executed and the id of the volume. This lock can then be used by other operations to avoid operation conflicts on shared volumes.
Example use:
If a volume operation uses this decorator, it will block until the named lock is free. This is used to protect concurrent operations on the same volume e.g. delete VolA while create volume VolB from VolA is in progress.
Drivers for volumes.
Bases: object
Executes commands relating to Volumes.
Base Driver for Cinder Volume Control Path, This includes supported/required implementation for API calls. Also provides generic implementation of core features like cloning, copy_image_to_volume etc, this way drivers that inherit from this base class and don’t offer their own impl can fall back on a general solution here.
Key thing to keep in mind with this driver is that it’s intended that these drivers ONLY implement Control Path details (create, delete, extend...), while transport or data path related implementation should be a member object that we call a connector. The point here is that for example don’t allow the LVM driver to implement iSCSI methods, instead call whatever connector it has configured via conf file (iSCSI{LIO, TGT, IET}, FC, etc).
In the base class and for example the LVM driver we do this via a has-a relationship and just provide an interface to the specific connector methods. How you do this in your own driver is of course up to you.
Callback for volume attached to instance or host.
Create a new backup from an existing volume.
Clean up after an interrupted image copy.
Fetch the image from image_service and write it to the volume.
Copy data from src_vol to dest_vol.
Copy the volume to the specified image.
Exports the volume.
Can optionally return a Dictionary of changes to the volume object to be persisted.
Creates a volume. Can optionally return a Dictionary of changes to the volume object to be persisted.
If volume_type extra specs includes ‘capabilities:replication <is> True’ the driver needs to create a volume replica (secondary), and setup replication between the newly created volume and the secondary volume. Returned dictionary should include:
volume[‘replication_status’] = ‘copying’ volume[‘replication_extended_status’] = driver specific value volume[‘driver_data’] = driver specific value
Deletes a volume.
If volume_type extra specs includes ‘replication: <is> True’ then the driver needs to delete the volume replica too.
Callback for volume detached.
Any initialization the volume driver does while starting.
Synchronously recreates an export for a volume.
Get the default filter_function string.
Each driver could overwrite the method to return a well-known default string if it is available.
Returns: | None |
---|
Get the default goodness_function string.
Each driver could overwrite the method to return a well-known default string if it is available.
Returns: | None |
---|
Get filter_function string.
Returns either the string from the driver instance or global section in cinder.conf. If nothing is specified in cinder.conf, then try to find the default filter_function. When None is returned the scheduler will always pass the driver instance.
:return a filter_function string or None
Get good_function string.
Returns either the string from the driver instance or global section in cinder.conf. If nothing is specified in cinder.conf, then try to find the default goodness_function. When None is returned the scheduler will give the lowest score to the driver instance.
:return a goodness_function string or None
Return pool name where volume reside on.
Parameters: | volume – The volume hosted by the the driver. |
---|---|
Returns: | name of the pool where given volume is in. |
Get the current version of this driver.
Return the current state of the volume service. If ‘refresh’ is True, run the update first.
For replication the following state should be reported: replication = True (None or false disables replication)
Allow connection to connector and return connection info.
Parameters: |
|
---|
connected to. :param initiator_data (optional): A dictionary of driver_initiator_data objects with key-value pairs that have been saved for this initiator by a driver in previous initialize_connection calls. :returns conn_info: A dictionary of connection information. This can optionally include a “initiator_updates” field.
The “initiator_updates” field must be a dictionary containing a “set_values” and/or “remove_values” field. The “set_values” field must be a dictionary of key-value pairs to be set/updated in the db. The “remove_values” field must be a list of keys, previously set with “set_values”, that will be deleted from the db.
Removes an export for a volume.
Restore an existing backup to a new or existing volume.
Determine if driver is running in Secure File Operations mode.
The Cinder Volume driver needs to query if this driver is running in a secure file operations mode. By default, it is False: any driver that does support secure file operations should override this method.
Disallow connection from connector
Fail if connector doesn’t contain all the data needed by driver.
Bases: object
Create a volume efficiently from an existing image.
image_location is a string whose format depends on the image service backend in use. The driver should use it to determine whether cloning is possible.
image_id is a string which represents id of the image. It can be used by the driver to introspect internal stores or registry to do an efficient image clone.
image_meta is a dictionary that includes ‘disk_format’ (e.g. raw, qcow2) and other image attributes that allow drivers to decide whether they can clone the image without first requiring conversion.
image_service is the reference of the image_service to use. Note that this is needed to be passed here for drivers that will want to fetch images from the image service directly.
Returns a dict of volume properties eg. provider_location, boolean indicating whether cloning occurred
Bases: object
Creates a clone of the specified volume.
If volume_type extra specs includes ‘replication: <is> True’ the driver needs to create a volume replica (secondary) and setup replication between the newly created volume and the secondary volume.
Bases: object
Creates a cgsnapshot.
Creates a consistencygroup.
Deletes a cgsnapshot.
Deletes a consistency group.
Bases: object
Bases: cinder.volume.driver.ISCSIDriver
Logs calls instead of executing.
No setup necessary in fake mode.
Creates a clone of the specified volume.
Exports the volume. Can optionally return a Dictionary of changes to the volume object to be persisted.
Creates a snapshot.
Creates a volume from a snapshot.
Deletes a snapshot.
Deletes a volume.
Synchronously recreates an export for a volume.
Execute that simply logs the command.
Removes an export for a volume.
Bases: cinder.volume.driver.FakeISCSIDriver
Logs calls instead of executing.
Execute that simply logs the command.
Bases: cinder.volume.driver.VolumeDriver
Executes commands relating to Fibre Channel volumes.
Get volume stats.
If ‘refresh’ is True, run update the stats first.
Initializes the connection and returns connection info.
The driver returns a driver_volume_type of ‘fibre_channel’. The target_wwn can be a single entry or a list of wwns that correspond to the list of remote wwn(s) that will export the volume. Example return values:
- {
‘driver_volume_type’: ‘fibre_channel’ ‘data’: {
‘target_discovered’: True, ‘target_lun’: 1, ‘target_wwn’: ‘1234567890123’, ‘access_mode’: ‘rw’}
}
or
- {
‘driver_volume_type’: ‘fibre_channel’ ‘data’: {
‘target_discovered’: True, ‘target_lun’: 1, ‘target_wwn’: [‘1234567890123’, ‘0987654321321’], ‘access_mode’: ‘rw’}
}
Fail if connector doesn’t contain all the data needed by driver.
Do a check on the connector and ensure that it has wwnns, wwpns.
Test for non-empty setting in connector.
Bases: cinder.volume.driver.VolumeDriver
Executes commands relating to ISCSI volumes.
We make use of model provider properties as follows:
Get volume stats.
If ‘refresh’ is True, run update the stats first.
Initializes the connection and returns connection info.
The iscsi driver returns a driver_volume_type of ‘iscsi’. The format of the driver data is defined in _get_iscsi_properties. Example return value:
{
'driver_volume_type': 'iscsi'
'data': {
'target_discovered': True,
'target_iqn': 'iqn.2010-10.org.openstack:volume-00000001',
'target_portal': '127.0.0.0.1:3260',
'volume_id': 1,
'access_mode': 'rw'
}
}
If the backend driver supports multiple connections for multipath and for single path with failover, “target_portals”, “target_iqns”, “target_luns” are also populated:
{
'driver_volume_type': 'iscsi'
'data': {
'target_discovered': False,
'target_iqn': 'iqn.2010-10.org.openstack:volume1',
'target_iqns': ['iqn.2010-10.org.openstack:volume1',
'iqn.2010-10.org.openstack:volume1-2'],
'target_portal': '10.0.0.1:3260',
'target_portals': ['10.0.0.1:3260', '10.0.1.1:3260']
'target_lun': 1,
'target_luns': [1, 1],
'volume_id': 1,
'access_mode': 'rw'
}
}
Bases: cinder.volume.driver.ISCSIDriver
Executes commands relating to ISER volumes.
We make use of model provider properties as follows:
Initializes the connection and returns connection info.
The iser driver returns a driver_volume_type of ‘iser’. The format of the driver data is defined in _get_iser_properties. Example return value:
{
'driver_volume_type': 'iser'
'data': {
'target_discovered': True,
'target_iqn':
'iqn.2010-10.org.iser.openstack:volume-00000001',
'target_portal': '127.0.0.0.1:3260',
'volume_id': 1,
}
}
Bases: object
Bases: object
Brings an existing backend storage object under Cinder management.
existing_ref is passed straight through from the API request’s manage_existing_ref value, and it is up to the driver how this should be interpreted. It should be sufficient to identify a storage object that the driver should somehow associate with the newly-created cinder volume structure.
There are two ways to do this:
If the existing_ref doesn’t make sense, or doesn’t refer to an existing backend storage object, raise a ManageExistingInvalidReference exception.
The volume may have a volume_type, and the driver can inspect that and compare against the properties of the referenced backend storage object. If they are incompatible, raise a ManageExistingVolumeTypeMismatch, specifying a reason for the failure.
Parameters: |
|
---|
volume
Return size of volume to be managed by manage_existing.
When calculating the size, round up to the next GB.
Parameters: |
|
---|
volume
Removes the specified volume from Cinder management.
Does not delete the underlying backend storage object.
For most drivers, this will not need to do anything. However, some drivers might use this call as an opportunity to clean up any Cinder-specific configuration that they have associated with the backend storage object.
Parameters: | volume – Cinder volume to unmanage |
---|
Bases: object
Migrate the volume to the specified host.
Returns a boolean indicating whether the migration occurred, as well as model_update.
Parameters: |
|
---|
Bases: object
Proxy Volume Driver to mark proxy drivers
If a driver uses a proxy class (e.g. by using __setattr__ and __getattr__) without directly inheriting from base volume driver this class can help marking them and retrieve the actual used driver object.
Bases: object
Creates a test replica clone of the specified replicated volume.
Create a clone of the replicated (secondary) volume.
Query the actual volume replication status from the driver.
Returns model_update for the volume. The driver is expected to update the following entries:
‘replication_status’ ‘replication_extended_status’ ‘replication_driver_data’
Possible ‘replication_status’ values (in model_update) are: ‘error’ - replication in error state ‘copying’ - replication copying data to secondary (inconsistent) ‘active’ - replication copying data to secondary (consistent) ‘active-stopped’ - replication data copy on hold (consistent) ‘inactive’ - replication data copy on hold (inconsistent) Values in ‘replication_extended_status’ and ‘replication_driver_data’ are managed by the driver.
Parameters: |
|
---|
Promote the replica to be the primary volume.
Following this command, replication between the volumes at the storage level should be stopped, the replica should be available to be attached, and the replication status should be in status ‘inactive’.
Returns model_update for the volume. The driver is expected to update the following entries:
‘replication_status’ ‘replication_extended_status’ ‘replication_driver_data’
Possible ‘replication_status’ values (in model_update) are: ‘error’ - replication in error state ‘inactive’ - replication data copy on hold (inconsistent) Values in ‘replication_extended_status’ and ‘replication_driver_data’ are managed by the driver.
Parameters: |
|
---|
Re-enable replication between the replica and primary volume.
This is used to re-enable/fix the replication between primary and secondary. One use is as part of the fail-back process, when you re-synchorize your old primary with the promoted volume (the old replica). Returns model_update for the volume to reflect the actions of the driver. The driver is expected to update the following entries:
‘replication_status’ ‘replication_extended_status’ ‘replication_driver_data’
Possible ‘replication_status’ values (in model_update) are: ‘error’ - replication in error state ‘copying’ - replication copying data to secondary (inconsistent) ‘active’ - replication copying data to secondary (consistent) ‘active-stopped’ - replication data copy on hold (consistent) ‘inactive’ - replication data copy on hold (inconsistent) Values in ‘replication_extended_status’ and ‘replication_driver_data’ are managed by the driver.
Parameters: |
|
---|
Bases: object
Convert the volume to be of the new type.
Returns either: A boolean indicating whether the retype occurred, or A tuple (retyped, model_update) where retyped is a boolean indicating if the retype occurred, and the model_update includes changes for the volume db. if diff[‘extra_specs’] includes ‘replication’ then:
- if (‘True’, _ ) then replication should be disabled:
- Volume replica should be deleted volume[‘replication_status’] should be changed to ‘disabled’ volume[‘replication_extended_status’] = None volume[‘replication_driver_data’] = None
- if (_, ‘True’) then replication should be enabled:
- Volume replica (secondary) should be created, and replication should be setup between the volume and the newly created replica volume[‘replication_status’] = ‘copying’ volume[‘replication_extended_status’] = driver specific value volume[‘replication_driver_data’] = driver specific value
Parameters: |
|
---|
Bases: object
Creates a snapshot.
Creates a volume from a snapshot.
If volume_type extra specs includes ‘replication: <is> True’ the driver needs to create a volume replica (secondary), and setup replication between the newly created volume and the secondary volume.
Deletes a snapshot.
Bases: object
Accept the transfer of a volume for a new user/project.
Bases: cinder.volume.driver.ConsistencyGroupVD, cinder.volume.driver.TransferVD, cinder.volume.driver.ManageableVD, cinder.volume.driver.ExtendVD, cinder.volume.driver.CloneableVD, cinder.volume.driver.CloneableImageVD, cinder.volume.driver.SnapshotVD, cinder.volume.driver.ReplicaVD, cinder.volume.driver.RetypeVD, cinder.volume.driver.LocalVD, cinder.volume.driver.MigrateVD, cinder.volume.driver.BaseVD
This class will be deprecated soon. Please us the abstract classes above for new drivers.
Creates a cgsnapshot.
Creates a consistencygroup.
Creates a consistencygroup from source.
Parameters: |
|
---|
:return model_update, volumes_model_update
Currently the source can only be cgsnapshot.
param volumes is retrieved directly from the db. It is a list of cinder.db.sqlalchemy.models.Volume to be precise. It cannot be assigned to volumes_model_update. volumes_model_update is a list of dictionaries. It has to be built by the driver. An entry will be in this format: [‘id’: xxx, ‘status’: xxx, ......]. model_update will be in this format: [‘status’: xxx, ......].
To be consistent with other volume operations, the manager will assume the operation is successful if no exception is thrown by the driver. For a successful operation, the driver can either build the model_update and volumes_model_update and return them or return None, None.
Deletes a cgsnapshot.
Deletes a consistency group.
Return pool name where volume reside on.
Parameters: | volume – The volume hosted by the the driver. |
---|---|
Returns: | name of the pool where given volume is in. |
Disallow connection from connector
Updates a consistency group.
Parameters: |
|
---|
:return model_update, add_volumes_update, remove_volumes_update
model_update is a dictionary that the driver wants the manager to update upon a successful return. If None is returned, the manager will set the status to ‘available’.
add_volumes_update and remove_volumes_update are lists of dictionaries that the driver wants the manager to update upon a successful return. Note that each entry requires a {‘id’: xxx} so that the correct volume entry can be updated. If None is returned, the volume will remain its original status. Also note that you cannot directly assign add_volumes to add_volumes_update as add_volumes is a list of cinder.db.sqlalchemy.models.Volume objects and cannot be used for db update directly. Same with remove_volumes.
If the driver throws an exception, the status of the group as well as those of the volumes to be added/removed will be set to ‘error’.
Return model update for migrated volume.
Parameters: |
|
---|
:return model_update to update DB with any needed changes
Cinder uses iSCSI to export storage volumes from multiple storage nodes. These iSCSI exports are attached (using libvirt) directly to running instances.
Cinder volumes are exported over the primary system VLAN (usually VLAN 1), and not over individual VLANs.
The underlying volumes by default are LVM logical volumes, created on demand within a single large volume group.