Image Live-Migration

RBD images can be live-migrated between different pools within the same cluster; between different image formats and layouts; or from external data sources. When started, the source will be deep-copied to the destination image, pulling all snapshot history while preserving the sparse allocation of data where possible.

By default, when live-migrating RBD images within the same Ceph cluster, the source image will be marked read-only and all clients will instead redirect IOs to the new target image. In addition, this mode can optionally preserve the link to the source image's parent to preserve sparseness, or it can flatten the image during the migration to remove the dependency on the source image's parent.

The live-migration process can also be used in an import-only mode where the source image remains unmodified and the target image can be linked to an external data source such as a backing file, HTTP(s) file, or S3 object.

The live-migration copy process can safely run in the background while the new target image is in use. There is currently a requirement to temporarily stop using the source image before preparing a migration when not using the import-only mode of operation. This helps to ensure that the client using the image is updated to point to the new target image.

备注

Image live-migration requires the Ceph Nautilus release or later. Support for external data sources requires the Ceph Pacific release of later. The krbd kernel module does not support live-migration at this time.

The live-migration process is comprised of three steps:

  1. Prepare Migration: The initial step creates the new target image and links the target image to the source. When not configured in the import-only mode, the source image will also be linked to the target image and marked read-only.

    Similar to layered images, attempts to read uninitialized data extents within the target image will internally redirect the read to the source image, and writes to uninitialized extents within the target will internally deep-copy the overlapping source image block to the target image.

  2. Execute Migration: This is a background operation that deep-copies all initialized blocks from the source image to the target. This step can be run while clients are actively using the new target image.

  3. Finish Migration: Once the background migration process has completed, the migration can be committed or aborted. Committing the migration will remove the cross-links between the source and target images, and will remove the source image if not configured in the import-only mode. Aborting the migration will remove the cross-links, and will remove the target image.

Prepare Migration

The default live-migration process for images within the same Ceph cluster is initiated by running the rbd migration prepare command, providing the source and target images:

$ rbd migration prepare migration_source [migration_target]

The rbd migration prepare command accepts all the same layout optionals as the rbd create command, which allows changes to the immutable image on-disk layout. The migration_target can be skipped if the goal is only to change the on-disk layout, keeping the original image name.

All clients using the source image must be stopped prior to preparing a live-migration. The prepare step will fail if it finds any running clients with the image open in read/write mode. Once the prepare step is complete, the clients can be restarted using the new target image name. Attempting to restart the clients using the source image name will result in failure.

The rbd status command will show the current state of the live-migration:

$ rbd status migration_target
Watchers: none
Migration:
            source: rbd/migration_source (5e2cba2f62e)
            destination: rbd/migration_target (5e2ed95ed806)
            state: prepared

Note that the source image will be moved to the RBD trash to avoid mistaken usage during the migration process:

$ rbd info migration_source
rbd: error opening image migration_source: (2) No such file or directory
$ rbd trash ls --all
5e2cba2f62e migration_source

Prepare Import-Only Migration

The import-only live-migration process is initiated by running the same rbd migration prepare command, but adding the --import-only optional and providing a JSON-encoded source-spec to describe how to access the source image data. This source-spec can either be passed directly via the --source-spec optional, or via a file or STDIN via the --source-spec-path optional:

$ rbd migration prepare --import-only --source-spec "<JSON>" migration_target

The rbd migration prepare command accepts all the same layout optionals as the rbd create command.

The rbd status command will show the current state of the live-migration:

$ rbd status migration_target
Watchers: none
Migration:
        source: {"stream":{"file_path":"/mnt/image.raw","type":"file"},"type":"raw"}
        destination: rbd/migration_target (ac69113dc1d7)
        state: prepared

The general format for the source-spec JSON is as follows:

{
    "type": "<format-type>",
    <format unique parameters>
    "stream": {
        "type": "<stream-type>",
        <stream unique parameters>
    }
}

The following formats are currently supported: native, qcow, and raw. The following streams are currently supported: file, http, and s3.

Formats

The native format can be used to describe a native RBD image within a Ceph cluster as the source image. Its source-spec JSON is encoded as follows:

{
    "type": "native",
    "pool_name": "<pool-name>",
    ["pool_id": <pool-id>,] (optional alternative to "pool_name")
    ["pool_namespace": "<pool-namespace",] (optional)
    "image_name": "<image-name>",
    ["image_id": "<image-id>",] (optional if image in trash)
    "snap_name": "<snap-name>",
    ["snap_id": "<snap-id>",] (optional alternative to "snap_name")
}

Note that the native format does not include the stream object since it utilizes native Ceph operations. For example, to import from the image rbd/ns1/image1@snap1, the source-spec could be encoded as:

{
    "type": "native",
    "pool_name": "rbd",
    "pool_namespace": "ns1",
    "image_name": "image1",
    "snap_name": "snap1"
}

The qcow format can be used to describe a QCOW (QEMU copy-on-write) block device. Both the QCOW (v1) and QCOW2 formats are currently supported with the exception of advanced features such as compression, encryption, backing files, and external data files. Support for these missing features may be added in a future release. The qcow format data can be linked to any supported stream source described below. For example, its base source-spec JSON is encoded as follows:

{
    "type": "qcow",
    "stream": {
        <stream unique parameters>
    }
}

The raw format can be used to describe a thick-provisioned, raw block device export (i.e. rbd export --export-format 1 <snap-spec>). The raw format data can be linked to any supported stream source described below. For example, its base source-spec JSON is encoded as follows:

{
    "type": "raw",
    "stream": {
        <stream unique parameters for HEAD, non-snapshot revision>
    },
    "snapshots": [
        {
            "type": "raw",
            "name": "<snapshot-name>",
            "stream": {
                <stream unique parameters for snapshot>
            }
        },
    ] (optional oldest to newest ordering of snapshots)
}

The inclusion of the snapshots array is optional and currently only supports thick-provisioned raw snapshot exports.

Additional formats such as RBD export-format v2 and RBD export-diff snapshots will be added in a future release.

Streams

The file stream can be used to import from a locally accessible POSIX file source. Its source-spec JSON is encoded as follows:

{
    <format unique parameters>
    "stream": {
        "type": "file",
        "file_path": "<file-path>"
    }
}

For example, to import a raw-format image from a file located at "/mnt/image.raw", its source-spec JSON is encoded as follows:

{
    "type": "raw",
    "stream": {
        "type": "file",
        "file_path": "/mnt/image.raw"
    }
}

The http stream can be used to import from a remote HTTP or HTTPS web server. Its source-spec JSON is encoded as follows:

{
    <format unique parameters>
    "stream": {
        "type": "http",
        "url": "<url-path>"
    }
}

For example, to import a raw-format image from a file located at http://download.ceph.com/image.raw, its source-spec JSON is encoded as follows:

{
    "type": "raw",
    "stream": {
        "type": "http",
        "url": "http://download.ceph.com/image.raw"
    }
}

The s3 stream can be used to import from a remote S3 bucket. Its source-spec JSON is encoded as follows:

{
    <format unique parameters>
    "stream": {
        "type": "s3",
        "url": "<url-path>",
        "access_key": "<access-key>",
        "secret_key": "<secret-key>"
    }
}

For example, to import a raw-format image from a file located at http://s3.ceph.com/bucket/image.raw, its source-spec JSON is encoded as follows:

{
    "type": "raw",
    "stream": {
        "type": "s3",
        "url": "http://s3.ceph.com/bucket/image.raw",
        "access_key": "NX5QOQKC6BH2IDN8HC7A",
        "secret_key": "LnEsqNNqZIpkzauboDcLXLcYaWwLQ3Kop0zAnKIn"
    }
}

备注

The access_key and secret_key parameters support storing the keys in the MON config-key store by prefixing the key values with config:// followed by the path in the MON config-key store to the value. Values can be stored in the config-key store via ceph config-key set <key-path> <value> (e.g. ceph config-key set rbd/s3/access_key NX5QOQKC6BH2IDN8HC7A).

Execute Migration

After preparing the live-migration, the image blocks from the source image must be copied to the target image. This is accomplished by running the rbd migration execute command:

$ rbd migration execute migration_target
Image migration: 100% complete...done.

The rbd status command will also provide feedback on the progress of the migration block deep-copy process:

$ rbd status migration_target
Watchers:
    watcher=1.2.3.4:0/3695551461 client.123 cookie=123
Migration:
            source: rbd/migration_source (5e2cba2f62e)
            destination: rbd/migration_target (5e2ed95ed806)
            state: executing (32% complete)

Commit Migration

Once the live-migration has completed deep-copying all data blocks from the source image to the target, the migration can be committed:

$ rbd status migration_target
Watchers: none
Migration:
            source: rbd/migration_source (5e2cba2f62e)
            destination: rbd/migration_target (5e2ed95ed806)
            state: executed
$ rbd migration commit migration_target
Commit image migration: 100% complete...done.

If the migration_source image is a parent of one or more clones, the --force option will need to be specified after ensuring all descendent clone images are not in use.

Committing the live-migration will remove the cross-links between the source and target images, and will remove the source image:

$ rbd trash list --all

Abort Migration

If you wish to revert the prepare or execute step, run the rbd migration abort command to revert the migration process:

$ rbd migration abort migration_target
Abort image migration: 100% complete...done.

Aborting the migration will result in the target image being deleted and access to the original source image being restored:

$ rbd ls
migration_source