manila/doc/source/admin/cephfs_driver.rst

CephFS driver

The CephFS driver enables manila to export shared filesystems backed by Ceph's File System (CephFS) using either the Ceph network protocol or NFS protocol. Guests require a native Ceph client or an NFS client in order to mount the filesystem.

When guests access CephFS using the native Ceph protocol, access is controlled via Ceph's cephx authentication system. If a user requests share access for an ID, Ceph creates a corresponding Ceph auth ID and a secret key if they do not already exist, and authorizes the ID to access the share. The client can then mount the share using the ID and the secret key. To learn more about configuring Ceph clients to access the shares created using this driver, please see the Ceph documentation

And when guests access CephFS through NFS, an NFS-Ganesha server mediates access to CephFS. The driver enables access control by managing the NFS-Ganesha server's exports.

Supported Operations

The following operations are supported with CephFS backend:

• Create/delete share
• Allow/deny access to share
  • Only cephx access type is supported for CephFS native protocol.
  • Only ip access type is supported for NFS protocol.
  • read-only and read-write access levels are supported.
• Extend/shrink share
• Create/delete snapshot
• Create/delete share groups
• Create/delete share group snapshots

Prerequisites

Important

A manila share backed by CephFS is only as good as the underlying filesystem. Take care when configuring your Ceph cluster, and consult the latest guidance on the use of CephFS in the Ceph documentation.

Ceph testing matrix

As Ceph and Manila continue to grow, it is essential to test and support combinations of releases supported by both projects. However, there is little community bandwidth to cover all of them. For simplicity sake, we are focused on testing (and therefore supporting) the current Ceph active releases. Check out the list of Ceph active releases here.

Below is the current state of testing for Ceph releases with this project. Adjacent components such as devstack-plugin-ceph and tripleo are added to the table below. Contributors to those projects can determine what versions of ceph are tested and supported with manila by those components; however, their state is presented here for ease of access.

Important

From the Victoria cycle, the Manila CephFS driver is not tested or supported with Ceph clusters older than Nautilus. Future releases of Manila may be incompatible with Nautilus too! We suggest always running the latest version of Manila with the latest release of Ceph.

OpenStack release	manila	devstack-plugin-ceph	tripleo
Queens	Luminous	Luminous	Luminous
Rocky	Luminous	Luminous	Luminous
Stein	Nautilus	Luminous, Nautilus	Nautilus
Train	Nautilus	Luminous, Nautilus	Nautilus
Ussuri	Nautilus	Luminous, Nautilus	Nautilus
Victoria	Nautilus	Nautilus, Octopus	Nautilus
Wallaby	Octopus	Nautilus, Octopus	Pacific

Additionally, it is expected that the version of the Ceph client available to manila is aligned with the Ceph server version. Mixing server and client versions is strongly unadvised.

In case of using the NFS Ganesha driver, it's also a good practice to use the versions that align with the Ceph version of choice.

Important

It's recommended to install the latest stable version of Ceph Nautilus/Octopus/Pacific release. See, Ceph releases

Prior to upgrading to Wallaby, please ensure that you're running at least the following versions of Ceph:

Release	Minimum version
Nautilus	14.2.20
Octopus	15.2.11
Pacific	16.2.1

Common Prerequisites

• A Ceph cluster with a filesystem configured (See Create ceph filesystem on how to create a filesystem.)
• python3-rados and python3-ceph-argparse packages installed in the servers running the manila-share service.
• Network connectivity between your Ceph cluster's public network and the servers running the manila-share service.

For CephFS native shares

• Ceph client installed in the guest
• Network connectivity between your Ceph cluster's public network and guests. See security_cephfs_native.

For CephFS NFS shares

• 3.0 or later versions of NFS-Ganesha.
• NFS client installed in the guest.
• Network connectivity between your Ceph cluster's public network and NFS-Ganesha server.
• Network connectivity between your NFS-Ganesha server and the manila guest.

Authorizing the driver to communicate with Ceph

Capabilities required for the Ceph manila identity have changed from the Wallaby release. The Ceph manila identity configured no longer needs any MDS capability. The MON and OSD capabilities can be reduced as well. However new MGR capabilities are now required. If not accorded, the driver cannot communicate to the Ceph Cluster.

Important

The driver in the Wallaby (or later) release requires a Ceph identity with a different set of Ceph capabilities when compared to the driver in a pre-Wallaby release.

When upgrading to Wallaby you'll also have to update the capabilities of the Ceph identity used by the driver (refer to Ceph user capabilities docs) E.g. a native driver that already uses client.manila Ceph identity, issue command ceph auth caps client.manila mon 'allow r' mgr 'allow rw'

For the CephFS Native driver, the auth ID should be set as follows:

ceph auth get-or-create client.manila -o manila.keyring \
  mgr 'allow rw' \
  mon 'allow r'

For the CephFS NFS driver, we use a specific pool to store exports (configurable with the config option "ganesha_rados_store_pool_name"). We also need to specify osd caps for it. So, the auth ID should be set as follows:

ceph auth get-or-create client.manila -o manila.keyring \
  osd 'allow rw pool=<ganesha_rados_store_pool_name>" \
  mgr 'allow rw' \
  mon 'allow r'

manila.keyring, along with your ceph.conf file, will then need to be placed on the server running the manila-share service.

Important

To communicate with the Ceph backend, a CephFS driver instance (represented as a backend driver section in manila.conf) requires its own Ceph auth ID that is not used by other CephFS driver instances running in the same controller node.

In the server running the manila-share service, you can place the ceph.conf and manila.keyring files in the /etc/ceph directory. Set the same owner for the manila-share process and the manila.keyring file. Add the following section to the ceph.conf file.

[client.manila]
client mount uid = 0
client mount gid = 0
log file = /opt/stack/logs/ceph-client.manila.log
admin socket = /opt/stack/status/stack/ceph-$name.$pid.asok
keyring = /etc/ceph/manila.keyring

It is advisable to modify the Ceph client's admin socket file and log file locations so that they are co-located with manila services's pid files and log files respectively.

Enabling snapshot support in Ceph backend

CephFS Snapshots were experimental prior to the Nautilus release of Ceph. There may be some limitations on snapshots based on the Ceph version you use.

From Ceph Nautilus, all new filesystems created on Ceph have snapshots enabled by default. If you've upgraded your ceph cluster and want to enable snapshots on a pre-existing filesystem, you can do so:

ceph fs set {fs_name} allow_new_snaps true

Configuring CephFS backend in manila.conf

Configure CephFS native share backend in manila.conf

Add CephFS to enabled_share_protocols (enforced at manila api layer). In this example we leave NFS and CIFS enabled, although you can remove these if you will only use a CephFS backend:

enabled_share_protocols = NFS,CIFS,CEPHFS

Create a section like this to define a CephFS native backend:

[cephfsnative1]
driver_handles_share_servers = False
share_backend_name = CEPHFSNATIVE1
share_driver = manila.share.drivers.cephfs.driver.CephFSDriver
cephfs_conf_path = /etc/ceph/ceph.conf
cephfs_protocol_helper_type = CEPHFS
cephfs_auth_id = manila
cephfs_cluster_name = ceph
cephfs_filesystem_name = cephfs

Set driver-handles-share-servers to False as the driver does not manage the lifecycle of share-servers. For the driver backend to expose shares via the native Ceph protocol, set cephfs_protocol_helper_type to CEPHFS.

Then edit enabled_share_backends to point to the driver's backend section using the section name. In this example we are also including another backend ("generic1"), you would include whatever other backends you have configured.

Finally, edit cephfs_filesystem_name with the name of the Ceph filesystem (also referred to as a CephFS volume) you want to use. If you have more than one Ceph filesystem in the cluster, you need to set this option.

enabled_share_backends = generic1, cephfsnative1

Configure CephFS NFS share backend in manila.conf

Note

Prior to configuring the Manila CephFS driver to use NFS, you must have installed and configured NFS-Ganesha. For guidance on configuration, refer to the NFS-Ganesha setup guide.

Add NFS to enabled_share_protocols if it's not already there:

enabled_share_protocols = NFS,CIFS,CEPHFS

Create a section to define a CephFS NFS share backend:

[cephfsnfs1]
driver_handles_share_servers = False
share_backend_name = CEPHFSNFS1
share_driver = manila.share.drivers.cephfs.driver.CephFSDriver
cephfs_protocol_helper_type = NFS
cephfs_conf_path = /etc/ceph/ceph.conf
cephfs_auth_id = manila
cephfs_cluster_name = ceph
cephfs_filesystem_name = cephfs
cephfs_ganesha_server_is_remote= False
cephfs_ganesha_server_ip = 172.24.4.3
ganesha_rados_store_enable = True
ganesha_rados_store_pool_name = cephfs_data

The following options are set in the driver backend section above:

• driver-handles-share-servers to False as the driver does not manage the lifecycle of share-servers.
• cephfs_protocol_helper_type to NFS to allow NFS protocol access to the CephFS backed shares.
• ceph_auth_id to the ceph auth ID created in authorize_ceph_driver.
• cephfs_ganesha_server_is_remote to False if the NFS-ganesha server is co-located with the manila-share service. If the NFS-Ganesha server is remote, then set the options to True, and set other options such as cephfs_ganesha_server_ip, cephfs_ganesha_server_username, and cephfs_ganesha_server_password (or cephfs_ganesha_path_to_private_key) to allow the driver to manage the NFS-Ganesha export entries over SSH.
• cephfs_ganesha_server_ip to the ganesha server IP address. It is recommended to set this option even if the ganesha server is co-located with the manila-share service.
• ganesha_rados_store_enable to True or False. Setting this option to True allows NFS Ganesha to store exports and its export counter in Ceph RADOS objects. We recommend setting this to True and using a RADOS object since it is useful for highly available NFS-Ganesha deployments to store their configuration efficiently in an already available distributed storage system.
• ganesha_rados_store_pool_name to the name of the RADOS pool you have created for use with NFS-Ganesha. Set this option only if also setting the ganesha_rados_store_enable option to True. If you want to use one of the backend CephFS's RADOS pools, then using CephFS's data pool is preferred over using its metadata pool.

Edit enabled_share_backends to point to the driver's backend section using the section name, cephfsnfs1.

Finally, edit cephfs_filesystem_name with the name of the Ceph filesystem (also referred to as a CephFS volume) you want to use. If you have more than one Ceph filesystem in the cluster, you need to set this option.

enabled_share_backends = generic1, cephfsnfs1

Space considerations

The CephFS driver reports total and free capacity available across the Ceph cluster to manila to allow provisioning. All CephFS shares are thinly provisioned, i.e., empty shares do not consume any significant space on the cluster. The CephFS driver does not allow controlling oversubscription via manila. So, as long as there is free space, provisioning will continue, and eventually this may cause your Ceph cluster to be over provisioned and you may run out of space if shares are being filled to capacity. It is advised that you use Ceph's monitoring tools to monitor space usage and add more storage when required in order to honor space requirements for provisioned manila shares. You may use the driver configuration option reserved_share_percentage to prevent manila from filling up your Ceph cluster, and allow existing shares to grow.

Creating shares

Create CephFS native share

The default share type may have driver_handles_share_servers set to True. Configure a share type suitable for CephFS native share:

manila type-create cephfsnativetype false
manila type-key cephfsnativetype set vendor_name=Ceph storage_protocol=CEPHFS

Then create a share,

manila create --share-type cephfsnativetype --name cephnativeshare1 cephfs 1

Note the export location of the share:

manila share-export-location-list cephnativeshare1

The export location of the share contains the Ceph monitor (mon) addresses and ports, and the path to be mounted. It is of the form, {mon ip addr:port}[,{mon ip addr:port}]:{path to be mounted}

Create CephFS NFS share

Configure a share type suitable for CephFS NFS share:

manila type-create cephfsnfstype false
manila type-key cephfsnfstype set vendor_name=Ceph storage_protocol=NFS

Then create a share:

manila create --share-type cephfsnfstype --name cephnfsshare1 nfs 1

Note the export location of the share:

manila share-export-location-list cephnfsshare1

The export location of the share contains the IP address of the NFS-Ganesha server and the path to be mounted. It is of the form, {NFS-Ganesha server address}:{path to be mounted}

Allowing access to shares

Allow access to CephFS native share

Allow Ceph auth ID alice access to the share using cephx access type.

manila access-allow cephnativeshare1 cephx alice

Note the access status, and the access/secret key of alice.

manila access-list cephnativeshare1

Allow access to CephFS NFS share

Allow a guest access to the share using ip access type.

manila access-allow cephnfsshare1 ip 172.24.4.225

Mounting CephFS shares

Mounting CephFS native share using FUSE client

Using the secret key of the authorized ID alice create a keyring file, alice.keyring like:

[client.alice]
        key = AQA8+ANW/4ZWNRAAOtWJMFPEihBA1unFImJczA==

Using the mon IP addresses from the share's export location, create a configuration file, ceph.conf like:

[client]
        client quota = true
        mon host = 192.168.1.7:6789, 192.168.1.8:6789, 192.168.1.9:6789

Finally, mount the filesystem, substituting the filenames of the keyring and configuration files you just created, and substituting the path to be mounted from the share's export location:

sudo ceph-fuse ~/mnt \
--id=alice \
--conf=./ceph.conf \
--keyring=./alice.keyring \
--client-mountpoint=/volumes/_nogroup/4c55ad20-9c55-4a5e-9233-8ac64566b98c

Mounting CephFS native share using Kernel client

If you have the ceph-common package installed in the client host, you can use the kernel client to mount CephFS shares.

Important

If you choose to use the kernel client rather than the FUSE client the share size limits set in manila may not be obeyed in versions of kernel older than 4.17 and Ceph versions older than mimic. See the quota limitations documentation to understand CephFS quotas.

The mount command is as follows:

mount -t ceph {mon1 ip addr}:6789,{mon2 ip addr}:6789,{mon3 ip addr}:6789:/ \
    {mount-point} -o name={access-id},secret={access-key}

With our earlier examples, this would be:

mount -t ceph 192.168.1.7:6789, 192.168.1.8:6789, 192.168.1.9:6789:/ \
    /volumes/_nogroup/4c55ad20-9c55-4a5e-9233-8ac64566b98c \
    -o name=alice,secret='AQA8+ANW/4ZWNRAAOtWJMFPEihBA1unFImJczA=='

Mount CephFS NFS share using NFS client

In the guest, mount the share using the NFS client and knowing the share's export location.

sudo mount -t nfs 172.24.4.3:/volumes/_nogroup/6732900b-32c1-4816-a529-4d6d3f15811e /mnt/nfs/

Known restrictions

• A CephFS driver instance, represented as a backend driver section in manila.conf, requires a Ceph auth ID unique to the backend Ceph Filesystem. Using a non-unique Ceph auth ID will result in the driver unintentionally evicting other CephFS clients using the same Ceph auth ID to connect to the backend.
• Snapshots are read-only. A user can read a snapshot's contents from the .snap/{manila-snapshot-id}_{unknown-id} folder within the mounted share.

Security

• Each share's data is mapped to a distinct Ceph RADOS namespace. A guest is restricted to access only that particular RADOS namespace. https://docs.ceph.com/docs/nautilus/cephfs/file-layouts/

• An additional level of resource isolation can be provided by mapping a share's contents to a separate RADOS pool. This layout would be preferred only for cloud deployments with a limited number of shares needing strong resource separation. You can do this by setting a share type specification, cephfs:data_isolated for the share type used by the cephfs driver.

  manila type-key cephfstype set cephfs:data_isolated=True

Security with CephFS native share backend

As the guests need direct access to Ceph's public network, CephFS native share backend is suitable only in private clouds where guests can be trusted.

The manila.share.drivers.cephfs.driver Module

manila.share.drivers.cephfs.driver