Copyright (C) 2016, Western Digital.
The dm-zoned device mapper target provides random write access to zoned block devices (ZBC and ZAC compliant devices). It hides to the device user (a file system or an application doing raw block device accesses) the sequential write constraint of host-managed zoned block devices, allowing the use of applications and file systems that do not have native zoned block device support.
File systems or applications that can natively support host-managed zoned block devices (e.g. the f2fs file system since kernel 4.10) do not need to use the dm-zoned device mapper target.
For a more detailed description of the zoned block device models and their constraints see the T10 ZBC specification for SCSI devices, and the T13 ZAC specification for ATA devices.
dm-zoned implementation focuses on simplicity and on minimizing resource usage overhead (CPU, memory and storage overhead). For a 10TB host-managed disk with 256 MB zones, dm-zoned memory usage per disk instance is at most 4.5 MB and as little as 5 zones will be used internally for storing metadata and performaing reclaim operations.
See the section dm-zoned Internals for more details.
The dmzadm utility allows formating backend devices for use with the dm-zoned device mapper target driver. This utility will inspect the device verifying that it is a zoned block device and will prepare and write on-disk dm-zoned metadata according to the device capacity, zone size, etc.
The dm-zoned-tools project source code is distributed under the terms of the GNU General Public License v3.0 or later (GPL-v3). A copy of this license with dm-zoned-tools copyright can be found in the files [LICENSES/GPL-3.0-or-later.txt] and [COPYING.GPL].
dm-zoned-tools and all its applications are distributed "as is", without technical support, and WITHOUT ANY WARRANTY, without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
All source files in dm-zoned-tools contain the GPL-3.0-or-later license SPDX short identifier in place of the full license text.
SPDX-License-Identifier: GPL-3.0-or-later
Some files such as the Makefile.am files and the .gitignore file are public
domain specified by the CC0 1.0 Universal (CC0 1.0) Public Domain Dedication.
These files are identified with the following SPDX header.
SPDX-License-Identifier: CC0-1.0
See the file [LICENSES/CC0-1.0.txt] for the full text of this license.
The following packages must be installed prior to compiling.
- pkg-config (pkgconf)
- m4
- autoconf
- automake
- libtool
- libuuid library and its development headers (libuuid and libuuid-devel packages)
- libblkid library and its development headers (libblkid and libblkid-devel packages)
- libudev library and its development headers (systemd and systemd-devel packages)
- device-mapper libraries and development headers (device-mapper-libs and device-mapper-devel packages)
- Linux kernel module management libraries and development headers (kmod and kmod-devel packages)
The following commands will compile the dmzadm tool.
$ sh ./autogen.sh
$ ./configure
$ make
To install the compiled dmzadm executable file, simply execute as root the following command.
$ make install
The default installation directory is /usr/sbin. This default location can be
changed using the configure script. Executing the following command displays
the options used to control the installation path.
$ ./configure --help
The rpm and rpmbuild utilities are necessary to build dm-zoned-tools RPM packages. Once these utilities are installed, the RPM packages can be built using the following command.
$ sh ./autogen.sh
$ ./configure
$ make rpm
Four RPM packages are built: a binary package providing dmmzadm executable and its documentation and license files, a source RPM package, a debuginfo RPM package and a debugsource RPM package.
The source RPM package can be used to build the binary and debug RPM packages outside of dm-zoned-tools source tree using the following command.
$ rpmbuild --rebuild dm-zoned-tools-<version>.src.rpm
The dm-zoned device mapper is included with the mainline Linux kernel code since kernel version 4.13.0. dm-zoned compilation must be enabled in the kernel configuration. This can be done by setting the CONFIG_DM_ZONED configuration option to "y" or "m" in the menu: Device Drivers -> Multiple devices driver support (RAID and LVM) -> Drive-managed zoned block device target support.
Since kernel 4.16.0, using the deadline or mq-deadline block I/O scheduler with zoned block devices is also necessary to avoid write request reordering leading to write I/O errors. A zoned block device must be setup with this scheduler before executing the dmzadm tool. This can be done using the following command.
# echo deadline > /sys/block/<disk name>/queue/scheduler
Alternatively, a udev rule can also be defined to force the use of the deadline scheduler on a particular disk or on all disk. An example of such rule is shown below.
ACTION=="add|change", KERNEL=="sd*[!0-9]", ATTRS{queue/zoned}=="host-managed", \
ATTR{queue/scheduler}="deadline"
dmzadm detailed usage is as follows.
> dmzadm --help
dmzadm allows formatting, checking and repairing
a zoned block device for use with the dm-zoned
device mapper.
Usage: dmzadm <operation> <device(s)> [options]
Operations
--version | -v : Print version number and exit
--help | -h : General help message
--format : Format a block device metadata
--check : Check a block device metadata
--repair : Repair a block device metadata
--relabel : Change the device label
--start : Start the device-mapper target
--stop : Stop the device-mapper target
Devices
For a single device target, a zoned block device
must be specified. For a multi-device target, a
a list of block devices must be specified, with
a regular block device as the first device specified,
followed by one or more zoned block devices
General options
--verbose : Verbose output
--vverbose : Very verbose output
Format operation options
--force : Force overwrite of existing content
--label=<str> : Set the target label name to <str>
--seq=<num> : Number of sequential zones reserved
for reclaim. The minimum is 1 and the
default is 16
Relabel operation options
--label=<str> : Set the target new label name to <str>
To create a dm-zoned target device, a zoned block device must first be formatted using the dmzadm tool. This tool will analyze the zone configuration of the device, determine where to place the metadata sets and initialize on disk the metadata used by the dm-zoned target driver.
Formatting a single device target is done using the command.
> dmzadm --format /dev/sdX
where /dev/sdX identifies the backend zoned block device to use.
Starting with Linux kernel v5.8.0, regular block devices such as SSDs can also be used together with zoned block devices with dm-zoned. In this case, conventional zones are emulated for the regular block device to hold dm-zoned metadata and for buffering data. When a regular block device is used, the zone reclaim process operates by copying data from emulated conventional zones on the regular block device to zones of the zoned block device. This dual-drive configuration can significantly increase performance of the target device under write-intensive workloads.
To format a dm-zoned target device using an additional regular block device and optionally several zoned block devices, the following commands can be used.
> dmzadm --format /dev/nvmeXnY /dev/sdZ /dev/sdZZ
Where /dev/nvmeXnY is in this example is a NVMe SSD and the scsi disks
/dev/sdZ and /dev/sdZZ` and zoned HDDs.
A formatted dm-zoned target device can be started by executing the following command.
> dmzadm --start /dev/sdX
For a multi-device target, the same list of devices as used for format must be specified.
> dmzadm --start /dev/nvmen0p1 /dev/sdX /dev/sdY
Conversely, a dm-zoned target device can be disabled using the --stop
operation.
> dmzadm --stop /dev/sdX
To report problems, please contact:
- Damien Le Moal (damien.lemoal@wdc.com)
PLEASE DO NOT SUBMIT CONFIDENTIAL INFORMATION OR INFORMATION SPECIFIC TO DRIVES THAT ARE VENDOR SAMPLES OR NOT PUBLICLY AVAILABLE.
Read the [CONTRIBUTING] file and send patches to:
Damien Le Moal <damien.lemoal@wdc.com>
Matias Bjørling <matias.bjorling@wdc.com>
If you believe your changes require kernel eyes or review, also Cc the device mapper kernel development mailing list at:
dm-devel@redhat.com
dm-zoned implements an on-disk write buffering scheme to handle random write accesses to sequential write required zones of a zoned block device. Conventional zones of the backend device are used for buffering random accesses, as well as for storing internal metadata.
Optionally, since Linux kernel version 5.8.0, an additional regular block device can also be used to provide randomly writable storage used in place of the conventional zones of the backend zoned block device for write buffering. With this new version of dm-zoned, multiple zoned block devices can also be used to increase performance.
All zones of the device(s) used to back a dm-zoned target are separated into 2 types:
-
Metadata zones: these are randomly writable zones used to store metadata. Randomly writable zones may be conventional zones or sequential write preferred zones (host-aware devices only). Metadata zones are not reported as usable capacity to the user. If an additional regular block device is used for write buffering, metadata zones are stored on this cache device.
-
Data zones: All remaining zones of the device. The majority of these zones will be sequential zones which are used used exclusively for storing user data. The conventional zones (or part of the sequential write preferred zones on a host-aware device) may be used also for buffering user random writes. User data may thus be stored either in conventional zone or in a sequential zone.
dm-zoned exposes a logical device with a sector size of 4096 bytes, irrespectively of the physical sector size of the backend zoned block device being used. This allows reducing the amount of metadata needed to manage valid blocks (blocks written). The on-disk metadata format is as follows:
-
The first block of the first randomly writable zone found contains the super block which describes the amount and position on disk of metadata blocks.
-
Following the super block, a set of blocks is used to describe the mapping of the logical chunks of the target logical device to data zones. The mapping is indexed by logical chunk number and each mapping entry indicates the data zone storing the chunk data and optionally the zone number of a random zone used to buffer random modification to the chunk data.
-
A set of blocks used to store bitmaps indicating the validity of blocks in the data zones follows the mapping table blocks. A valid block is a block that was writen and not discarded. For a buffered data zone, a block can be valid only in the data zone or in the buffer zone.
For a logical chunk mapped to a random data zone, all write operations are processed by directly writing to the data zone. If the mapping zone is to a sequential zone, the write operation is processed directly only and only if the write offset within the logical chunk is equal to the write pointer offset within of the sequential data zone (i.e. the write operation is aligned on the zone write pointer). Otherwise, write operations are processed indirectly using a buffer zone: a randomly writable free data zone is allocated and assigned to the chunk being accessed in addition to the already mapped sequential data zone. Writing block to the buffer zone will invalidate the same blocks in the sequential data zone.
Read operations are processed according to the block validity information provided by the bitmaps: valid blocks are read either from the data zone or, if the data zone is buffered, from the buffer zone assigned to the data zone.
After some time, the limited number of random zones available may be exhausted and unaligned writes to unbuffered zones become impossible. To avoid such situation, a reclaim process regularly scans used random zones and try to "reclaim" them by copying (sequentially) the valid blocks of the buffer zone to a free sequential zone. Once the copy completes, the chunk mapping is updated to point to the sequential zone and the buffer zone freed for reuse.
To protect internal metadata against corruption in case of sudden power loss or system crash, 2 sets of metadata zones are used. One set, the primary set, is used as the main metadata repository, while the secondary set is used as a log. Modified metadata are first written to the secondary set and the log so created validated by writing an updated super block in the secondary set. Once this log operation completes, updates in place of metadata blocks can be done in the primary metadata set, ensuring that one of the set is always correct. Flush operations are used as a commit point: upon reception of a flush operation, metadata activity is temporarily stopped, all dirty metadata logged and updated and normal operation resumed. This only temporarily delays write and discard requests. Read requests can be processed while metadata logging is executed.