1 ========
2 dm-zoned
3 ========
4
5 The dm-zoned device mapper target exposes a zo
6 ZAC compliant devices) as a regular block devi
7 pattern constraints. In effect, it implements
8 block device which hides from the user (a file
9 doing raw block device accesses) the sequentia
10 host-managed zoned block devices and can mitig
11 device-side performance degradation due to exc
12 host-aware zoned block devices.
13
14 For a more detailed description of the zoned b
15 their constraints see (for SCSI devices):
16
17 https://www.t10.org/drafts.htm#ZBC_Family
18
19 and (for ATA devices):
20
21 http://www.t13.org/Documents/UploadedDocuments
22
23 The dm-zoned implementation is simple and mini
24 and memory usage as well as storage capacity l
25 host-managed disk with 256 MB zones, dm-zoned
26 instance is at most 4.5 MB and as little as 5
27 internally for storing metadata and performing
28
29 dm-zoned target devices are formatted and chec
30 utility available at:
31
32 https://github.com/hgst/dm-zoned-tools
33
34 Algorithm
35 =========
36
37 dm-zoned implements an on-disk buffering schem
38 write accesses to the sequential zones of a zo
39 Conventional zones are used for caching as wel
40 metadata. It can also use a regular block devi
41 block device; in that case the regular block d
42 in zones with the same size as the zoned block
43 placed in front of the zones from the zoned bl
44 just like conventional zones.
45
46 The zones of the device(s) are separated into
47
48 1) Metadata zones: these are conventional zone
49 Metadata zones are not reported as usable capa
50
51 2) Data zones: all remaining zones, the vast m
52 sequential zones used exclusively to store use
53 zones of the device may be used also for buffe
54 Data in these zones may be directly mapped to
55 later moved to a sequential zone so that the c
56 reused for buffering incoming random writes.
57
58 dm-zoned exposes a logical device with a secto
59 irrespective of the physical sector size of th
60 device being used. This allows reducing the am
61 manage valid blocks (blocks written).
62
63 The on-disk metadata format is as follows:
64
65 1) The first block of the first conventional z
66 super block which describes the on disk amount
67 blocks.
68
69 2) Following the super block, a set of blocks
70 mapping of the logical device blocks. The mapp
71 blocks, with the chunk size equal to the zoned
72 mapping table is indexed by chunk number and e
73 indicates the zone number of the device storin
74 mapping entry may also indicate if the zone nu
75 zone used to buffer random modification to the
76
77 3) A set of blocks used to store bitmaps indic
78 blocks in the data zones follows the mapping t
79 defined as a block that was written and not di
80 data chunk, a block is always valid only in th
81 chunk or in the buffer zone of the chunk.
82
83 For a logical chunk mapped to a conventional z
84 are processed by directly writing to the zone.
85 sequential zone, the write operation is proces
86 write offset within the logical chunk is equal
87 offset within of the sequential data zone (i.e
88 aligned on the zone write pointer). Otherwise,
89 processed indirectly using a buffer zone. In t
90 conventional zone is allocated and assigned to
91 accessed. Writing a block to the buffer zone o
92 automatically invalidate the same block in the
93 the chunk. If all blocks of the sequential zon
94 is freed and the chunk buffer zone becomes the
95 chunk, resulting in native random write perfor
96 block device.
97
98 Read operations are processed according to the
99 information provided by the bitmaps. Valid blo
100 the sequential zone mapping a chunk, or if the
101 the buffer zone assigned. If the accessed chun
102 accessed blocks are invalid, the read buffer i
103 operation terminated.
104
105 After some time, the limited number of convent
106 be exhausted (all used to map chunks or buffer
107 unaligned writes to unbuffered chunks become i
108 situation, a reclaim process regularly scans u
109 tries to reclaim the least recently used zones
110 blocks of the buffer zone to a free sequential
111 completes, the chunk mapping is updated to poi
112 and the buffer zone freed for reuse.
113
114 Metadata Protection
115 ===================
116
117 To protect metadata against corruption in case
118 system crash, 2 sets of metadata zones are use
119 set, is used as the main metadata region, whil
120 used as a staging area. Modified metadata is f
121 secondary set and validated by updating the su
122 set, a generation counter is used to indicate
123 newest metadata. Once this operation completes
124 block updates can be done in the primary metad
125 one of the set is always consistent (all modif
126 at all). Flush operations are used as a commit
127 a flush request, metadata modification activit
128 (for both incoming BIO processing and reclaim
129 metadata blocks are staged and updated. Normal
130 resumed. Flushing metadata thus only temporari
131 discard requests. Read requests can be process
132 metadata flush is being executed.
133
134 If a regular device is used in conjunction wit
135 a third set of metadata (without the zone bitm
136 start of the zoned block device. This metadata
137 '0' and will never be updated during normal op
138 identification purposes. The first and second
139 are located at the start of the regular block
140
141 Usage
142 =====
143
144 A zoned block device must first be formatted u
145 will analyze the device zone configuration, de
146 metadata sets on the device and initialize the
147
148 Ex::
149
150 dmzadm --format /dev/sdxx
151
152
153 If two drives are to be used, both devices mus
154 regular block device as the first device.
155
156 Ex::
157
158 dmzadm --format /dev/sdxx /dev/sdyy
159
160
161 Formatted device(s) can be started with the dm
162
163 Ex::
164
165 dmzadm --start /dev/sdxx /dev/sdyy
166
167
168 Information about the internal layout and curr
169 be obtained with the 'status' callback from dm
170
171 Ex::
172
173 dmsetup status /dev/dm-X
174
175 will return a line
176
177 0 <size> zoned <nr_zones> zones <nr_un
178
179 where <nr_zones> is the total number of zones,
180 of unmapped (ie free) random zones, <nr_rnd> t
181 <nr_unmap_seq> the number of unmapped sequenti
182 total number of sequential zones.
183
184 Normally the reclaim process will be started o
185 percent free random zones. In order to start t
186 even before reaching this threshold the 'dmset
187 used:
188
189 Ex::
190
191 dmsetup message /dev/dm-X 0 reclaim
192
193 will start the reclaim process and random zone
194 zones.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.