[PATCH v2 0/3] qapi & doc: deprecate drive-backup
Hi all! See 03 commit message for details. 01-02 are preparation docs update. v2: add a lot of documentation changes v1 was "[PATCH] qapi: deprecate drive-backup" Supersedes: <20210423125900.3640-1-vsementsov@virtuozzo.com> Vladimir Sementsov-Ogievskiy (3): docs/block-replication: use blockdev-backup docs/interop/bitmaps: use blockdev-backup qapi: deprecate drive-backup docs/block-replication.txt | 4 +- docs/interop/bitmaps.rst | 285 +++++++++++++++++++------ docs/interop/live-block-operations.rst | 47 ++-- docs/system/deprecated.rst | 11 + qapi/block-core.json | 5 +- 5 files changed, 263 insertions(+), 89 deletions(-) -- 2.29.2
We are going to deprecate drive-backup, so don't mention it here. Moreover, blockdev-backup seems more correct in the context. Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> --- docs/block-replication.txt | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/block-replication.txt b/docs/block-replication.txt index 108e9166a8..59eb2b33b3 100644 --- a/docs/block-replication.txt +++ b/docs/block-replication.txt @@ -79,7 +79,7 @@ Primary | || Secondary disk <--------- hidden-disk 5 <--------- || | | || | | || '-------------------------' - || drive-backup sync=none 6 + || blockdev-backup sync=none 6 1) The disk on the primary is represented by a block device with two children, providing replication between a primary disk and the host that @@ -101,7 +101,7 @@ should support bdrv_make_empty() and backing file. that is modified by the primary VM. It should also start as an empty disk, and the driver supports bdrv_make_empty() and backing file. -6) The drive-backup job (sync=none) is run to allow hidden-disk to buffer +6) The blockdev-backup job (sync=none) is run to allow hidden-disk to buffer any state that would otherwise be lost by the speculative write-through of the NBD server into the secondary disk. So before block replication, the primary disk and secondary disk should contain the same data. -- 2.29.2
On 5/5/21 9:58 AM, Vladimir Sementsov-Ogievskiy wrote:
We are going to deprecate drive-backup, so don't mention it here. Moreover, blockdev-backup seems more correct in the context.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: John Snow <jsnow@redhat.com>
--- docs/block-replication.txt | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/docs/block-replication.txt b/docs/block-replication.txt index 108e9166a8..59eb2b33b3 100644 --- a/docs/block-replication.txt +++ b/docs/block-replication.txt @@ -79,7 +79,7 @@ Primary | || Secondary disk <--------- hidden-disk 5 <--------- || | | || | | || '-------------------------' - || drive-backup sync=none 6 + || blockdev-backup sync=none 6
1) The disk on the primary is represented by a block device with two children, providing replication between a primary disk and the host that @@ -101,7 +101,7 @@ should support bdrv_make_empty() and backing file. that is modified by the primary VM. It should also start as an empty disk, and the driver supports bdrv_make_empty() and backing file.
-6) The drive-backup job (sync=none) is run to allow hidden-disk to buffer +6) The blockdev-backup job (sync=none) is run to allow hidden-disk to buffer any state that would otherwise be lost by the speculative write-through of the NBD server into the secondary disk. So before block replication, the primary disk and secondary disk should contain the same data.
We are going to deprecate drive-backup, so use modern interface here. In examples where target image creation is shown, show blockdev-add as well. If target creation omitted, omit blockdev-add as well. Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> --- docs/interop/bitmaps.rst | 285 +++++++++++++++++++++++++++++---------- 1 file changed, 215 insertions(+), 70 deletions(-) diff --git a/docs/interop/bitmaps.rst b/docs/interop/bitmaps.rst index 059ad67929..ef95090c81 100644 --- a/docs/interop/bitmaps.rst +++ b/docs/interop/bitmaps.rst @@ -539,12 +539,11 @@ other partial disk images on top of a base image to reconstruct a full backup from the point in time at which the incremental backup was issued. The "Push Model" here references the fact that QEMU is "pushing" the modified -blocks out to a destination. We will be using the `drive-backup -<qemu-qmp-ref.html#index-drive_002dbackup>`_ and `blockdev-backup -<qemu-qmp-ref.html#index-blockdev_002dbackup>`_ QMP commands to create both +blocks out to a destination. We will be using the `blockdev-backup +<qemu-qmp-ref.html#index-blockdev_002dbackup>`_ QMP command to create both full and incremental backups. -Both of these commands are jobs, which have their own QMP API for querying and +The command is a job, which has its own QMP API for querying and management documented in `Background jobs <qemu-qmp-ref.html#Background-jobs>`_. @@ -557,6 +556,10 @@ create a new incremental backup chain attached to a drive. This example creates a new, full backup of "drive0" and accompanies it with a new, empty bitmap that records writes from this point in time forward. +The target may be created with help of `blockdev-add +<qemu-qmp-ref.html#index-blockdev_002dadd>`_ or `blockdev-create +<qemu-qmp-ref.html#index-blockdev_002dcreate>`_ command. + .. note:: Any new writes that happen after this command is issued, even while the backup job runs, will be written locally and not to the backup destination. These writes will be recorded in the bitmap @@ -576,12 +579,11 @@ new, empty bitmap that records writes from this point in time forward. } }, { - "type": "drive-backup", + "type": "blockdev-backup", "data": { "device": "drive0", - "target": "/path/to/drive0.full.qcow2", - "sync": "full", - "format": "qcow2" + "target": "target0", + "sync": "full" } } ] @@ -664,12 +666,11 @@ use a transaction to reset the bitmap while making a new full backup: } }, { - "type": "drive-backup", + "type": "blockdev-backup", "data": { "device": "drive0", - "target": "/path/to/drive0.new_full.qcow2", - "sync": "full", - "format": "qcow2" + "target": "target0", + "sync": "full" } } ] @@ -728,19 +729,35 @@ Example: First Incremental Backup $ qemu-img create -f qcow2 drive0.inc0.qcow2 \ -b drive0.full.qcow2 -F qcow2 +#. Add target block node: + + .. code-block:: QMP + + -> { + "execute": "blockdev-add", + "arguments": { + "node-name": "target0", + "driver": "qcow2", + "file": { + "driver": "file", + "filename": "drive0.inc0.qcow2" + } + } + } + + <- { "return": {} } + #. Issue an incremental backup command: .. code-block:: QMP -> { - "execute": "drive-backup", + "execute": "blockdev-backup", "arguments": { "device": "drive0", "bitmap": "bitmap0", - "target": "drive0.inc0.qcow2", - "format": "qcow2", - "sync": "incremental", - "mode": "existing" + "target": "target0", + "sync": "incremental" } } @@ -785,20 +802,36 @@ Example: Second Incremental Backup $ qemu-img create -f qcow2 drive0.inc1.qcow2 \ -b drive0.inc0.qcow2 -F qcow2 +#. Add target block node: + + .. code-block:: QMP + + -> { + "execute": "blockdev-add", + "arguments": { + "node-name": "target0", + "driver": "qcow2", + "file": { + "driver": "file", + "filename": "drive0.inc1.qcow2" + } + } + } + + <- { "return": {} } + #. Issue a new incremental backup command. The only difference here is that we have changed the target image below. .. code-block:: QMP -> { - "execute": "drive-backup", + "execute": "blockdev-backup", "arguments": { "device": "drive0", "bitmap": "bitmap0", - "target": "drive0.inc1.qcow2", - "format": "qcow2", - "sync": "incremental", - "mode": "existing" + "target": "target0", + "sync": "incremental" } } @@ -866,20 +899,36 @@ image: file for you, but you lose control over format options like compatibility and preallocation presets. +#. Add target block node: + + .. code-block:: QMP + + -> { + "execute": "blockdev-add", + "arguments": { + "node-name": "target0", + "driver": "qcow2", + "file": { + "driver": "file", + "filename": "drive0.inc2.qcow2" + } + } + } + + <- { "return": {} } + #. Issue a new incremental backup command. Apart from the new destination image, there is no difference from the last two examples. .. code-block:: QMP -> { - "execute": "drive-backup", + "execute": "blockdev-backup", "arguments": { "device": "drive0", "bitmap": "bitmap0", - "target": "drive0.inc2.qcow2", - "format": "qcow2", - "sync": "incremental", - "mode": "existing" + "target": "target0", + "sync": "incremental" } } @@ -930,6 +979,38 @@ point in time. $ qemu-img create -f qcow2 drive0.full.qcow2 64G $ qemu-img create -f qcow2 drive1.full.qcow2 64G +#. Add target block nodes: + + .. code-block:: QMP + + -> { + "execute": "blockdev-add", + "arguments": { + "node-name": "target0", + "driver": "qcow2", + "file": { + "driver": "file", + "filename": "drive0.full.qcow2" + } + } + } + + <- { "return": {} } + + -> { + "execute": "blockdev-add", + "arguments": { + "node-name": "target1", + "driver": "qcow2", + "file": { + "driver": "file", + "filename": "drive1.full.qcow2" + } + } + } + + <- { "return": {} } + #. Create a full (anchor) backup for each drive, with accompanying bitmaps: .. code-block:: QMP @@ -953,21 +1034,19 @@ point in time. } }, { - "type": "drive-backup", + "type": "blockdev-backup", "data": { "device": "drive0", - "target": "/path/to/drive0.full.qcow2", - "sync": "full", - "format": "qcow2" + "target": "target0", + "sync": "full" } }, { - "type": "drive-backup", + "type": "blockdev-backup", "data": { "device": "drive1", - "target": "/path/to/drive1.full.qcow2", - "sync": "full", - "format": "qcow2" + "target": "target1", + "sync": "full" } } ] @@ -1016,6 +1095,38 @@ point in time. $ qemu-img create -f qcow2 drive1.inc0.qcow2 \ -b drive1.full.qcow2 -F qcow2 +#. Add target block nodes: + + .. code-block:: QMP + + -> { + "execute": "blockdev-add", + "arguments": { + "node-name": "target0", + "driver": "qcow2", + "file": { + "driver": "file", + "filename": "drive0.inc0.qcow2" + } + } + } + + <- { "return": {} } + + -> { + "execute": "blockdev-add", + "arguments": { + "node-name": "target1", + "driver": "qcow2", + "file": { + "driver": "file", + "filename": "drive1.inc0.qcow2" + } + } + } + + <- { "return": {} } + #. Issue a multi-drive incremental push backup transaction: .. code-block:: QMP @@ -1025,25 +1136,21 @@ point in time. "arguments": { "actions": [ { - "type": "drive-backup", + "type": "blockev-backup", "data": { "device": "drive0", "bitmap": "bitmap0", - "format": "qcow2", - "mode": "existing", "sync": "incremental", - "target": "drive0.inc0.qcow2" + "target": "target0" } }, { - "type": "drive-backup", + "type": "blockdev-backup", "data": { "device": "drive1", "bitmap": "bitmap0", - "format": "qcow2", - "mode": "existing", "sync": "incremental", - "target": "drive1.inc0.qcow2" + "target": "target1" } }, ] @@ -1119,19 +1226,35 @@ described above. This example demonstrates the single-job failure case: $ qemu-img create -f qcow2 drive0.inc0.qcow2 \ -b drive0.full.qcow2 -F qcow2 +#. Add target block node: + + .. code-block:: QMP + + -> { + "execute": "blockdev-add", + "arguments": { + "node-name": "target0", + "driver": "qcow2", + "file": { + "driver": "file", + "filename": "drive0.inc0.qcow2" + } + } + } + + <- { "return": {} } + #. Attempt to create an incremental backup via QMP: .. code-block:: QMP -> { - "execute": "drive-backup", + "execute": "blockdev-backup", "arguments": { "device": "drive0", "bitmap": "bitmap0", - "target": "drive0.inc0.qcow2", - "format": "qcow2", - "sync": "incremental", - "mode": "existing" + "target": "target0", + "sync": "incremental" } } @@ -1164,6 +1287,19 @@ described above. This example demonstrates the single-job failure case: "event": "BLOCK_JOB_COMPLETED" } +#. Remove target node: + + .. code-block:: QMP + + -> { + "execute": "blockdev-del", + "arguments": { + "node-name": "target0", + } + } + + <- { "return": {} } + #. Delete the failed image, and re-create it. .. code:: bash @@ -1172,20 +1308,36 @@ described above. This example demonstrates the single-job failure case: $ qemu-img create -f qcow2 drive0.inc0.qcow2 \ -b drive0.full.qcow2 -F qcow2 +#. Add target block node: + + .. code-block:: QMP + + -> { + "execute": "blockdev-add", + "arguments": { + "node-name": "target0", + "driver": "qcow2", + "file": { + "driver": "file", + "filename": "drive0.inc0.qcow2" + } + } + } + + <- { "return": {} } + #. Retry the command after fixing the underlying problem, such as freeing up space on the backup volume: .. code-block:: QMP -> { - "execute": "drive-backup", + "execute": "blockdev-backup", "arguments": { "device": "drive0", "bitmap": "bitmap0", - "target": "drive0.inc0.qcow2", - "format": "qcow2", - "sync": "incremental", - "mode": "existing" + "target": "target0", + "sync": "incremental" } } @@ -1210,7 +1362,8 @@ described above. This example demonstrates the single-job failure case: Example: Partial Transactional Failures ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -QMP commands like `drive-backup <qemu-qmp-ref.html#index-drive_002dbackup>`_ +QMP commands like `blockdev-backup +<qemu-qmp-ref.html#index-blockdev_002dbackup>`_ conceptually only start a job, and so transactions containing these commands may succeed even if the job it created later fails. This might have surprising interactions with notions of how a "transaction" ought to behave. @@ -1240,25 +1393,21 @@ and one succeeds: "arguments": { "actions": [ { - "type": "drive-backup", + "type": "blockdev-backup", "data": { "device": "drive0", "bitmap": "bitmap0", - "format": "qcow2", - "mode": "existing", "sync": "incremental", - "target": "drive0.inc0.qcow2" + "target": "target0" } }, { - "type": "drive-backup", + "type": "blockdev-backup", "data": { "device": "drive1", "bitmap": "bitmap0", - "format": "qcow2", - "mode": "existing", "sync": "incremental", - "target": "drive1.inc0.qcow2" + "target": "target1" } }] } @@ -1375,25 +1524,21 @@ applied: }, "actions": [ { - "type": "drive-backup", + "type": "blockdev-backup", "data": { "device": "drive0", "bitmap": "bitmap0", - "format": "qcow2", - "mode": "existing", "sync": "incremental", - "target": "drive0.inc0.qcow2" + "target": "target0" } }, { - "type": "drive-backup", + "type": "blockdev-backup", "data": { "device": "drive1", "bitmap": "bitmap0", - "format": "qcow2", - "mode": "existing", "sync": "incremental", - "target": "drive1.inc0.qcow2" + "target": "target1" } }] } -- 2.29.2
On Wed, May 05, 2021 at 04:58:02PM +0300, Vladimir Sementsov-Ogievskiy wrote:
We are going to deprecate drive-backup, so use modern interface here. In examples where target image creation is shown, show blockdev-add as well. If target creation omitted, omit blockdev-add as well.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> --- docs/interop/bitmaps.rst | 285 +++++++++++++++++++++++++++++---------- 1 file changed, 215 insertions(+), 70 deletions(-)
Looks fine to me. I have a couple of nits below, perhaps whoever is applying the patch can tweak them. FWIW: Reviewed-by: Kashyap Chamarthy <kchamart@redhat.com>
diff --git a/docs/interop/bitmaps.rst b/docs/interop/bitmaps.rst index 059ad67929..ef95090c81 100644 --- a/docs/interop/bitmaps.rst +++ b/docs/interop/bitmaps.rst @@ -539,12 +539,11 @@ other partial disk images on top of a base image to reconstruct a full backup from the point in time at which the incremental backup was issued.
The "Push Model" here references the fact that QEMU is "pushing" the modified -blocks out to a destination. We will be using the `drive-backup -<qemu-qmp-ref.html#index-drive_002dbackup>`_ and `blockdev-backup -<qemu-qmp-ref.html#index-blockdev_002dbackup>`_ QMP commands to create both +blocks out to a destination. We will be using the `blockdev-backup +<qemu-qmp-ref.html#index-blockdev_002dbackup>`_ QMP command to create both full and incremental backups.
-Both of these commands are jobs, which have their own QMP API for querying and +The command is a job, which has its own QMP API for querying and
nit: s/job/background job/
management documented in `Background jobs <qemu-qmp-ref.html#Background-jobs>`_.
@@ -557,6 +556,10 @@ create a new incremental backup chain attached to a drive. This example creates a new, full backup of "drive0" and accompanies it with a new, empty bitmap that records writes from this point in time forward.
+The target may be created with help of `blockdev-add
It is less ambiguous (at least to my eyes) to replace "may" with "can". nit: s/with help of/with the help of/
+<qemu-qmp-ref.html#index-blockdev_002dadd>`_ or `blockdev-create +<qemu-qmp-ref.html#index-blockdev_002dcreate>`_ command. + .. note:: Any new writes that happen after this command is issued, even while the backup job runs, will be written locally and not to the backup destination. These writes will be recorded in the bitmap
[...] -- /kashyap
On 5/5/21 9:58 AM, Vladimir Sementsov-Ogievskiy wrote:
We are going to deprecate drive-backup, so use modern interface here. In examples where target image creation is shown, show blockdev-add as well. If target creation omitted, omit blockdev-add as well.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Seems good, thanks! (aside: I really need to push forward with the QMP cross-references work ...)
Modern way is using blockdev-add + blockdev-backup, which provides a lot more control on how target is opened. As example of drive-backup problems consider the following: User of drive-backup expects that target will be opened in the same cache and aio mode as source. Corresponding logic is in drive_backup_prepare(), where we take bs->open_flags of source. It works rather bad if source was added by blockdev-add. Assume source is qcow2 image. On blockdev-add we should specify aio and cache options for file child of qcow2 node. What happens next: drive_backup_prepare() looks at bs->open_flags of qcow2 source node. But there no BDRV_O_NOCAHE neither BDRV_O_NATIVE_AIO: BDRV_O_NOCAHE is places in bs->file->bs->open_flags, and BDRV_O_NATIVE_AIO is nowhere, as file-posix parse options and simply set s->use_linux_aio. The documentation is updated in a minimal way, so that drive-backup is noted only as a deprecated command, and blockdev-backup used in most of places. Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> --- TODO: We also need to deprecate drive-backup transaction action.. But union members in QAPI doesn't support 'deprecated' feature. I tried to dig a bit, but failed :/ Markus, could you please help with it? At least by advice? docs/interop/live-block-operations.rst | 47 +++++++++++++++++--------- docs/system/deprecated.rst | 11 ++++++ qapi/block-core.json | 5 ++- 3 files changed, 46 insertions(+), 17 deletions(-) diff --git a/docs/interop/live-block-operations.rst b/docs/interop/live-block-operations.rst index 1073b930dc..f71f79ae2a 100644 --- a/docs/interop/live-block-operations.rst +++ b/docs/interop/live-block-operations.rst @@ -116,8 +116,8 @@ QEMU block layer supports. (3) ``drive-mirror`` (and ``blockdev-mirror``): Synchronize a running disk to another image. -(4) ``drive-backup`` (and ``blockdev-backup``): Point-in-time (live) copy - of a block device to a destination. +(4) ``blockdev-backup`` (and deprecated ``drive-backup``): Point-in-time + (live) copy of a block device to a destination. .. _`Interacting with a QEMU instance`: @@ -553,13 +553,14 @@ Currently, there are four different kinds: (3) ``none`` -- Synchronize only the new writes from this point on. - .. note:: In the case of ``drive-backup`` (or ``blockdev-backup``), - the behavior of ``none`` synchronization mode is different. - Normally, a ``backup`` job consists of two parts: Anything - that is overwritten by the guest is first copied out to - the backup, and in the background the whole image is - copied from start to end. With ``sync=none``, it's only - the first part. + .. note:: In the case of ``blockdev-backup`` (or deprecated + ``drive-backup``), the behavior of ``none`` + synchronization mode is different. Normally, a + ``backup`` job consists of two parts: Anything that is + overwritten by the guest is first copied out to the + backup, and in the background the whole image is copied + from start to end. With ``sync=none``, it's only the + first part. (4) ``incremental`` -- Synchronize content that is described by the dirty bitmap @@ -924,19 +925,22 @@ Shutdown the guest, by issuing the ``quit`` QMP command:: } -Live disk backup --- ``drive-backup`` and ``blockdev-backup`` -------------------------------------------------------------- +Live disk backup --- ``blockdev-backup`` and deprecated``drive-backup`` +----------------------------------------------------------------------- -The ``drive-backup`` (and its newer equivalent ``blockdev-backup``) allows +The ``blockdev-backup`` (and deprecated ``drive-backup``) allows you to create a point-in-time snapshot. -In this case, the point-in-time is when you *start* the ``drive-backup`` -(or its newer equivalent ``blockdev-backup``) command. +In this case, the point-in-time is when you *start* the +``blockdev-backup`` (or deprecated ``drive-backup``) command. QMP invocation for ``drive-backup`` ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +Note that ``drive-backup`` command is deprecated since Qemu 6.1 and +will be removed in future. + Yet again, starting afresh with our example disk image chain:: [A] <-- [B] <-- [C] <-- [D] @@ -961,11 +965,22 @@ will be issued, indicating the live block device job operation has completed, and no further action is required. +Moving from deprecated ``drive-backup`` to newer ``blockdev-backup`` +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +``blockdev-backup`` differs in a way of specifying backup target. +With ``blockdev-backup`` you can't specify filename as a target. +Instead you use node-name of existing block node, which you may add +by ``blockdev-add`` or ``blockdev-create`` commands. Correspondingly, +``blockdev-backup`` doesn't have ``mode`` and ``format`` arguments +which don't apply to existing block node. See following sections for +details and examples. + + Notes on ``blockdev-backup`` ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -The ``blockdev-backup`` command is equivalent in functionality to -``drive-backup``, except that it operates at node-level in a Block Driver +The ``blockdev-backup`` operates at node-level in a Block Driver State (BDS) graph. E.g. the sequence of actions to create a point-in-time backup diff --git a/docs/system/deprecated.rst b/docs/system/deprecated.rst index 80cae86252..676d72a1ed 100644 --- a/docs/system/deprecated.rst +++ b/docs/system/deprecated.rst @@ -186,6 +186,17 @@ Use the more generic commands ``block-export-add`` and ``block-export-del`` instead. As part of this deprecation, where ``nbd-server-add`` used a single ``bitmap``, the new ``block-export-add`` uses a list of ``bitmaps``. +``drive-backup`` (since 6.1) +'''''''''''''''''''''''''''' + +Use ``blockdev-backup`` in pair with ``blockdev-add`` instead. +This change primarily separates the creation/opening process of the backup +target with explicit, separate steps. ``blockdev-backup`` uses mostly the +same arguments as ``drive-backup``, except the ``format`` and ``mode`` +options are removed in favor of using explicit ``blockdev-create`` and +``blockdev-add`` calls. See :doc:`/interop/live-block-operations` for +details. + System accelerators ------------------- diff --git a/qapi/block-core.json b/qapi/block-core.json index 6d227924d0..8e2c6e1622 100644 --- a/qapi/block-core.json +++ b/qapi/block-core.json @@ -1642,6 +1642,9 @@ # The operation can be stopped before it has completed using the # block-job-cancel command. # +# Features: +# @deprecated: This command is deprecated. Use @blockdev-backup instead. +# # Returns: - nothing on success # - If @device is not a valid block device, GenericError # @@ -1657,7 +1660,7 @@ # ## { 'command': 'drive-backup', 'boxed': true, - 'data': 'DriveBackup' } + 'data': 'DriveBackup', 'features': ['deprecated'] } ## # @blockdev-backup: -- 2.29.2
On Wed, May 05, 2021 at 04:58:03PM +0300, Vladimir Sementsov-Ogievskiy wrote:
Modern way is using blockdev-add + blockdev-backup, which provides a lot more control on how target is opened.
As example of drive-backup problems consider the following:
User of drive-backup expects that target will be opened in the same cache and aio mode as source. Corresponding logic is in drive_backup_prepare(), where we take bs->open_flags of source.
It works rather bad if source was added by blockdev-add. Assume source is qcow2 image. On blockdev-add we should specify aio and cache options for file child of qcow2 node. What happens next:
drive_backup_prepare() looks at bs->open_flags of qcow2 source node. But there no BDRV_O_NOCAHE neither BDRV_O_NATIVE_AIO: BDRV_O_NOCAHE is places in bs->file->bs->open_flags, and BDRV_O_NATIVE_AIO is nowhere, as file-posix parse options and simply set s->use_linux_aio.
The documentation is updated in a minimal way, so that drive-backup is noted only as a deprecated command, and blockdev-backup used in most of places.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> ---
TODO: We also need to deprecate drive-backup transaction action.. But union members in QAPI doesn't support 'deprecated' feature. I tried to dig a bit, but failed :/ Markus, could you please help with it? At least by advice?
docs/interop/live-block-operations.rst | 47 +++++++++++++++++--------- docs/system/deprecated.rst | 11 ++++++ qapi/block-core.json | 5 ++-
The core changes itself looks good; I have some minor nit-picks below, hope that's not annoying. :-) With those addressed: Reviewed-by: Kashyap Chamarthy <kchamart@redhat.com>
3 files changed, 46 insertions(+), 17 deletions(-)
diff --git a/docs/interop/live-block-operations.rst b/docs/interop/live-block-operations.rst index 1073b930dc..f71f79ae2a 100644 --- a/docs/interop/live-block-operations.rst +++ b/docs/interop/live-block-operations.rst @@ -116,8 +116,8 @@ QEMU block layer supports. (3) ``drive-mirror`` (and ``blockdev-mirror``): Synchronize a running disk to another image.
-(4) ``drive-backup`` (and ``blockdev-backup``): Point-in-time (live) copy - of a block device to a destination. +(4) ``blockdev-backup`` (and deprecated ``drive-backup``): Point-in-time + (live) copy of a block device to a destination.
nit: s/deprecated ``drive-backup``/the deprecated ``drive-backup``/
.. _`Interacting with a QEMU instance`: @@ -553,13 +553,14 @@ Currently, there are four different kinds:
(3) ``none`` -- Synchronize only the new writes from this point on.
- .. note:: In the case of ``drive-backup`` (or ``blockdev-backup``), - the behavior of ``none`` synchronization mode is different. - Normally, a ``backup`` job consists of two parts: Anything - that is overwritten by the guest is first copied out to - the backup, and in the background the whole image is - copied from start to end. With ``sync=none``, it's only - the first part. + .. note:: In the case of ``blockdev-backup`` (or deprecated + ``drive-backup``), the behavior of ``none`` + synchronization mode is different. Normally, a + ``backup`` job consists of two parts: Anything that is + overwritten by the guest is first copied out to the + backup, and in the background the whole image is copied + from start to end. With ``sync=none``, it's only the + first part.
(4) ``incremental`` -- Synchronize content that is described by the dirty bitmap @@ -924,19 +925,22 @@ Shutdown the guest, by issuing the ``quit`` QMP command:: }
-Live disk backup --- ``drive-backup`` and ``blockdev-backup`` -------------------------------------------------------------- +Live disk backup --- ``blockdev-backup`` and deprecated``drive-backup`` +-----------------------------------------------------------------------
Here too, missing the article "the": "the deprecated".
-The ``drive-backup`` (and its newer equivalent ``blockdev-backup``) allows +The ``blockdev-backup`` (and deprecated ``drive-backup``) allows you to create a point-in-time snapshot.
-In this case, the point-in-time is when you *start* the ``drive-backup`` -(or its newer equivalent ``blockdev-backup``) command. +In this case, the point-in-time is when you *start* the +``blockdev-backup`` (or deprecated ``drive-backup``) command.
QMP invocation for ``drive-backup`` ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+Note that ``drive-backup`` command is deprecated since Qemu 6.1 and +will be removed in future.
nit: Let's consistently spell QEMU in all caps, please: s/Qemu/QEMU/
Yet again, starting afresh with our example disk image chain::
[A] <-- [B] <-- [C] <-- [D] @@ -961,11 +965,22 @@ will be issued, indicating the live block device job operation has completed, and no further action is required.
+Moving from deprecated ``drive-backup`` to newer ``blockdev-backup`` +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
s/from/from the/
+``blockdev-backup`` differs in a way of specifying backup target.
I might slightly rephrase it this way: ``blockdev-backup`` differs from ``drive-backup`` in how you specify the backup target.
+With ``blockdev-backup`` you can't specify filename as a target. +Instead you use node-name of existing block node, which you may add
Can use literals also for node-name: s/node-name/``node-name``
+by ``blockdev-add`` or ``blockdev-create`` commands. Correspondingly, +``blockdev-backup`` doesn't have ``mode`` and ``format`` arguments +which don't apply to existing block node. See following sections for
s/to/to an/
+details and examples. + + Notes on ``blockdev-backup`` ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-The ``blockdev-backup`` command is equivalent in functionality to -``drive-backup``, except that it operates at node-level in a Block Driver +The ``blockdev-backup`` operates at node-level in a Block Driver State (BDS) graph.
s/``blockdev-backup``/``blockdev-backup`` command/
E.g. the sequence of actions to create a point-in-time backup diff --git a/docs/system/deprecated.rst b/docs/system/deprecated.rst index 80cae86252..676d72a1ed 100644 --- a/docs/system/deprecated.rst +++ b/docs/system/deprecated.rst @@ -186,6 +186,17 @@ Use the more generic commands ``block-export-add`` and ``block-export-del`` instead. As part of this deprecation, where ``nbd-server-add`` used a single ``bitmap``, the new ``block-export-add`` uses a list of ``bitmaps``.
+``drive-backup`` (since 6.1) +'''''''''''''''''''''''''''' + +Use ``blockdev-backup`` in pair with ``blockdev-add`` instead.
nit: s/in pair/in combination/
+This change primarily separates the creation/opening process of the backup +target with explicit, separate steps. ``blockdev-backup`` uses mostly the +same arguments as ``drive-backup``, except the ``format`` and ``mode`` +options are removed in favor of using explicit ``blockdev-create`` and +``blockdev-add`` calls. See :doc:`/interop/live-block-operations` for +details.
The rest reads good to me. Thanks for fixing this. [...] -- /kashyap
On 5/6/21 5:57 AM, Kashyap Chamarthy wrote:
TODO: We also need to deprecate drive-backup transaction action.. But union members in QAPI doesn't support 'deprecated' feature. I tried to dig a bit, but failed :/ Markus, could you please help with it? At least by advice?
Oho, I see. OK, I'm not Markus, but I've been getting into lots of trouble in the QAPI generator lately, so let me see if I can help get you started... https://gitlab.com/jsnow/qemu/-/commits/hack-deprecate-union-branches/ Here's a quick hack that might expose that feature. I suppose we can discuss this with Markus and turn these into real patches if that's the direction we wanna head. --js
15.05.2021 01:38, John Snow wrote:
On 5/6/21 5:57 AM, Kashyap Chamarthy wrote:
TODO: We also need to deprecate drive-backup transaction action.. But union members in QAPI doesn't support 'deprecated' feature. I tried to dig a bit, but failed :/ Markus, could you please help with it? At least by advice?
Oho, I see.
OK, I'm not Markus, but I've been getting into lots of trouble in the QAPI generator lately, so let me see if I can help get you started...
https://gitlab.com/jsnow/qemu/-/commits/hack-deprecate-union-branches/
Here's a quick hack that might expose that feature. I suppose we can discuss this with Markus and turn these into real patches if that's the direction we wanna head.
Hi! Markus is silent.. Maybe, you'll send patches ? ) -- Best regards, Vladimir
On 5/24/21 10:06 AM, Vladimir Sementsov-Ogievskiy wrote:
15.05.2021 01:38, John Snow wrote:
On 5/6/21 5:57 AM, Kashyap Chamarthy wrote:
TODO: We also need to deprecate drive-backup transaction action.. But union members in QAPI doesn't support 'deprecated' feature. I tried to dig a bit, but failed :/ Markus, could you please help with it? At least by advice?
Oho, I see.
OK, I'm not Markus, but I've been getting into lots of trouble in the QAPI generator lately, so let me see if I can help get you started...
https://gitlab.com/jsnow/qemu/-/commits/hack-deprecate-union-branches/
Here's a quick hack that might expose that feature. I suppose we can discuss this with Markus and turn these into real patches if that's the direction we wanna head.
Hi! Markus is silent.. Maybe, you'll send patches ? )
He just went on PTO for 2 weeks :') It's going to have to wait, I'm afraid ...
24.05.2021 21:37, John Snow wrote:
On 5/24/21 10:06 AM, Vladimir Sementsov-Ogievskiy wrote:
15.05.2021 01:38, John Snow wrote:
On 5/6/21 5:57 AM, Kashyap Chamarthy wrote:
TODO: We also need to deprecate drive-backup transaction action.. But union members in QAPI doesn't support 'deprecated' feature. I tried to dig a bit, but failed :/ Markus, could you please help with it? At least by advice?
Oho, I see.
OK, I'm not Markus, but I've been getting into lots of trouble in the QAPI generator lately, so let me see if I can help get you started...
https://gitlab.com/jsnow/qemu/-/commits/hack-deprecate-union-branches/
Here's a quick hack that might expose that feature. I suppose we can discuss this with Markus and turn these into real patches if that's the direction we wanna head.
Hi! Markus is silent.. Maybe, you'll send patches ? )
He just went on PTO for 2 weeks :')
It's going to have to wait, I'm afraid ...
Hi! Any plans or updates? John, may be you just send your patches? -- Best regards, Vladimir
Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> writes:
24.05.2021 21:37, John Snow wrote:
On 5/24/21 10:06 AM, Vladimir Sementsov-Ogievskiy wrote:
15.05.2021 01:38, John Snow wrote:
On 5/6/21 5:57 AM, Kashyap Chamarthy wrote:
TODO: We also need to deprecate drive-backup transaction action.. But union members in QAPI doesn't support 'deprecated' feature. I tried to dig a bit, but failed :/ Markus, could you please help with it? At least by advice?
Oho, I see.
OK, I'm not Markus, but I've been getting into lots of trouble in the QAPI generator lately, so let me see if I can help get you started...
https://gitlab.com/jsnow/qemu/-/commits/hack-deprecate-union-branches/
Here's a quick hack that might expose that feature. I suppose we can discuss this with Markus and turn these into real patches if that's the direction we wanna head.
Hi! Markus is silent.. Maybe, you'll send patches ? )
He just went on PTO for 2 weeks :') It's going to have to wait, I'm afraid ...
Hi!
Any plans or updates? John, may be you just send your patches?
Yes, please! I intend to resume working on feature flags next week.
Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> writes: [...]
TODO: We also need to deprecate drive-backup transaction action.. But union members in QAPI doesn't support 'deprecated' feature. I tried to dig a bit, but failed :/ Markus, could you please help with it? At least by advice?
There are two closely related things in play here: the union branch and the corresponding enum value. So far, the QAPI schema language doesn't support tacking feature flags to either. If an enum value is deprecated, any union branches corresponding to it must also be deprecated (because their use requires using the deprecated enum value). The converse is not true, but I can't see a use for deprecating a union branch without also deprecating the enum member. I think we can implement feature flags just for enum members, then document that 'deprecated' enum value implies corresponding union branches are also deprecated. I have unfinished patches implementing feature flags for enum members. Since TransactionAction is a simple union, the corresponding enum is implicit. We can make it explicit by converting to a flat union. Simple unions need to die anyway. Does this make sense?
08.06.2021 14:12, Markus Armbruster wrote:
Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> writes:
[...]
TODO: We also need to deprecate drive-backup transaction action.. But union members in QAPI doesn't support 'deprecated' feature. I tried to dig a bit, but failed :/ Markus, could you please help with it? At least by advice?
There are two closely related things in play here: the union branch and the corresponding enum value.
So far, the QAPI schema language doesn't support tacking feature flags to either.
If an enum value is deprecated, any union branches corresponding to it must also be deprecated (because their use requires using the deprecated enum value).
The converse is not true, but I can't see a use for deprecating a union branch without also deprecating the enum member.
I think we can implement feature flags just for enum members, then document that 'deprecated' enum value implies corresponding union branches are also deprecated.
I have unfinished patches implementing feature flags for enum members.
Since TransactionAction is a simple union, the corresponding enum is implicit. We can make it explicit by converting to a flat union. Simple unions need to die anyway.
Does BlockStatsSpecific from qapi/block-core.json a correct example of flat union you mean? I can make patch to convert TransactionAction to be similar if that helps (discriminator field should be called "type", yes?).
Does this make sense?
Yes if it helps) Did you also look at John's https://gitlab.com/jsnow/qemu/-/commits/hack-deprecate-union-branches/ ? I hope you and John will send patches that you have, I'll help with reviewing (keep me in CC), and finally we'll get the feature. -- Best regards, Vladimir
Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> writes:
08.06.2021 14:12, Markus Armbruster wrote:
Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> writes:
[...]
TODO: We also need to deprecate drive-backup transaction action.. But union members in QAPI doesn't support 'deprecated' feature. I tried to dig a bit, but failed :/ Markus, could you please help with it? At least by advice?
There are two closely related things in play here: the union branch and the corresponding enum value.
So far, the QAPI schema language doesn't support tacking feature flags to either.
If an enum value is deprecated, any union branches corresponding to it must also be deprecated (because their use requires using the deprecated enum value).
The converse is not true, but I can't see a use for deprecating a union branch without also deprecating the enum member.
I think we can implement feature flags just for enum members, then document that 'deprecated' enum value implies corresponding union branches are also deprecated.
I have unfinished patches implementing feature flags for enum members.
Since TransactionAction is a simple union, the corresponding enum is implicit. We can make it explicit by converting to a flat union. Simple unions need to die anyway.
Does BlockStatsSpecific from qapi/block-core.json a correct example of flat union you mean? I can make patch to convert TransactionAction to be similar if that helps (discriminator field should be called "type", yes?).
From docs/devel/qapi-code-gen.txt:
A simple union can always be re-written as a flat union where the base class has a single member named 'type', and where each branch of the union has a struct with a single member named 'data'. That is, { 'union': 'Simple', 'data': { 'one': 'str', 'two': 'int' } } is identical on the wire to: { 'enum': 'Enum', 'data': ['one', 'two'] } { 'struct': 'Branch1', 'data': { 'data': 'str' } } { 'struct': 'Branch2', 'data': { 'data': 'int' } } { 'union': 'Flat', 'base': { 'type': 'Enum' }, 'discriminator': 'type', 'data': { 'one': 'Branch1', 'two': 'Branch2' } } The generated C isn't identical, but adjusting the code using it should be straightforward.
Does this make sense?
Yes if it helps)
Did you also look at John's https://gitlab.com/jsnow/qemu/-/commits/hack-deprecate-union-branches/ ?
Not yet.
I hope you and John will send patches that you have, I'll help with reviewing (keep me in CC), and finally we'll get the feature.
Sounds like a plan. I need to get my post-vacation e-mail pileup under control first.
09.06.2021 13:49, Markus Armbruster wrote:
Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> writes:
08.06.2021 14:12, Markus Armbruster wrote:
Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> writes:
[...]
TODO: We also need to deprecate drive-backup transaction action.. But union members in QAPI doesn't support 'deprecated' feature. I tried to dig a bit, but failed :/ Markus, could you please help with it? At least by advice?
There are two closely related things in play here: the union branch and the corresponding enum value.
So far, the QAPI schema language doesn't support tacking feature flags to either.
If an enum value is deprecated, any union branches corresponding to it must also be deprecated (because their use requires using the deprecated enum value).
The converse is not true, but I can't see a use for deprecating a union branch without also deprecating the enum member.
I think we can implement feature flags just for enum members, then document that 'deprecated' enum value implies corresponding union branches are also deprecated.
I have unfinished patches implementing feature flags for enum members.
Since TransactionAction is a simple union, the corresponding enum is implicit. We can make it explicit by converting to a flat union. Simple unions need to die anyway.
Does BlockStatsSpecific from qapi/block-core.json a correct example of flat union you mean? I can make patch to convert TransactionAction to be similar if that helps (discriminator field should be called "type", yes?).
From docs/devel/qapi-code-gen.txt:
A simple union can always be re-written as a flat union where the base class has a single member named 'type', and where each branch of the union has a struct with a single member named 'data'. That is,
{ 'union': 'Simple', 'data': { 'one': 'str', 'two': 'int' } }
is identical on the wire to:
{ 'enum': 'Enum', 'data': ['one', 'two'] } { 'struct': 'Branch1', 'data': { 'data': 'str' } } { 'struct': 'Branch2', 'data': { 'data': 'int' } } { 'union': 'Flat', 'base': { 'type': 'Enum' }, 'discriminator': 'type', 'data': { 'one': 'Branch1', 'two': 'Branch2' } }
The generated C isn't identical, but adjusting the code using it should be straightforward.
Does this make sense?
Yes if it helps)
Did you also look at John's https://gitlab.com/jsnow/qemu/-/commits/hack-deprecate-union-branches/ ?
Not yet.
I hope you and John will send patches that you have, I'll help with reviewing (keep me in CC), and finally we'll get the feature.
Sounds like a plan. I need to get my post-vacation e-mail pileup under control first.
Hi! Kindly remind in the case you forget :) Or may be I miss some patches? -- Best regards, Vladimir
Markus Armbruster <armbru@redhat.com> writes:
Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> writes:
08.06.2021 14:12, Markus Armbruster wrote:
Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> writes:
[...]
TODO: We also need to deprecate drive-backup transaction action.. But union members in QAPI doesn't support 'deprecated' feature. I tried to dig a bit, but failed :/ Markus, could you please help with it? At least by advice?
There are two closely related things in play here: the union branch and the corresponding enum value.
So far, the QAPI schema language doesn't support tacking feature flags to either.
If an enum value is deprecated, any union branches corresponding to it must also be deprecated (because their use requires using the deprecated enum value).
The converse is not true, but I can't see a use for deprecating a union branch without also deprecating the enum member.
I think we can implement feature flags just for enum members, then document that 'deprecated' enum value implies corresponding union branches are also deprecated.
I have unfinished patches implementing feature flags for enum members.
Since TransactionAction is a simple union, the corresponding enum is implicit. We can make it explicit by converting to a flat union. Simple unions need to die anyway.
Does BlockStatsSpecific from qapi/block-core.json a correct example of flat union you mean? I can make patch to convert TransactionAction to be similar if that helps (discriminator field should be called "type", yes?).
From docs/devel/qapi-code-gen.txt:
A simple union can always be re-written as a flat union where the base class has a single member named 'type', and where each branch of the union has a struct with a single member named 'data'. That is,
{ 'union': 'Simple', 'data': { 'one': 'str', 'two': 'int' } }
is identical on the wire to:
{ 'enum': 'Enum', 'data': ['one', 'two'] } { 'struct': 'Branch1', 'data': { 'data': 'str' } } { 'struct': 'Branch2', 'data': { 'data': 'int' } } { 'union': 'Flat', 'base': { 'type': 'Enum' }, 'discriminator': 'type', 'data': { 'one': 'Branch1', 'two': 'Branch2' } }
The generated C isn't identical, but adjusting the code using it should be straightforward.
Does this make sense?
Yes if it helps)
Did you also look at John's https://gitlab.com/jsnow/qemu/-/commits/hack-deprecate-union-branches/ ?
Not yet.
I hope you and John will send patches that you have, I'll help with reviewing (keep me in CC), and finally we'll get the feature.
Sounds like a plan. I need to get my post-vacation e-mail pileup under control first.
Just sent: Subject: [PATCH RFC 0/5] Subject: [PATCH RFC 0/5] qapi: Add feature flags to enum members Message-Id: <20210915192425.4104210-1-armbru@redhat.com> Yes, I mangled the subject %-/
participants (4)
-
John Snow -
Kashyap Chamarthy -
Markus Armbruster -
Vladimir Sementsov-Ogievskiy