1 ================= 1 =================
2 Writecache target 2 Writecache target
3 ================= 3 =================
4 4
5 The writecache target caches writes on persist 5 The writecache target caches writes on persistent memory or on SSD. It
6 doesn't cache reads because reads are supposed 6 doesn't cache reads because reads are supposed to be cached in page cache
7 in normal RAM. 7 in normal RAM.
8 8
9 When the device is constructed, the first sect 9 When the device is constructed, the first sector should be zeroed or the
10 first sector should contain valid superblock f 10 first sector should contain valid superblock from previous invocation.
11 11
12 Constructor parameters: 12 Constructor parameters:
13 13
14 1. type of the cache device - "p" or "s" 14 1. type of the cache device - "p" or "s"
15 - p - persistent memory 15 - p - persistent memory
16 - s - SSD 16 - s - SSD
17 2. the underlying device that will be cached 17 2. the underlying device that will be cached
18 3. the cache device 18 3. the cache device
19 4. block size (4096 is recommended; the maximu 19 4. block size (4096 is recommended; the maximum block size is the page
20 size) 20 size)
21 5. the number of optional parameters (the para 21 5. the number of optional parameters (the parameters with an argument
22 count as two) 22 count as two)
23 23
24 start_sector n (default: 0) 24 start_sector n (default: 0)
25 offset from the start of cache 25 offset from the start of cache device in 512-byte sectors
26 high_watermark n (default: 50) 26 high_watermark n (default: 50)
27 start writeback when the numbe 27 start writeback when the number of used blocks reach this
28 watermark 28 watermark
29 low_watermark x (default: 45) 29 low_watermark x (default: 45)
30 stop writeback when the number 30 stop writeback when the number of used blocks drops below
31 this watermark 31 this watermark
32 writeback_jobs n (default: unli 32 writeback_jobs n (default: unlimited)
33 limit the number of blocks tha 33 limit the number of blocks that are in flight during
34 writeback. Setting this value 34 writeback. Setting this value reduces writeback
35 throughput, but it may improve 35 throughput, but it may improve latency of read requests
36 autocommit_blocks n (default: 64 f 36 autocommit_blocks n (default: 64 for pmem, 65536 for ssd)
37 when the application writes th 37 when the application writes this amount of blocks without
38 issuing the FLUSH request, the 38 issuing the FLUSH request, the blocks are automatically
39 committed 39 committed
40 autocommit_time ms (default: 1000 40 autocommit_time ms (default: 1000)
41 autocommit time in millisecond 41 autocommit time in milliseconds. The data is automatically
42 committed if this time passes 42 committed if this time passes and no FLUSH request is
43 received 43 received
44 fua (by default on 44 fua (by default on)
45 applicable only to persistent 45 applicable only to persistent memory - use the FUA flag
46 when writing data from persist 46 when writing data from persistent memory back to the
47 underlying device 47 underlying device
48 nofua 48 nofua
49 applicable only to persistent 49 applicable only to persistent memory - don't use the FUA
50 flag when writing back data an 50 flag when writing back data and send the FLUSH request
51 afterwards 51 afterwards
52 52
53 - some underlying devices perf 53 - some underlying devices perform better with fua, some
54 with nofua. The user should 54 with nofua. The user should test it
55 cleaner 55 cleaner
56 when this option is activated 56 when this option is activated (either in the constructor
57 arguments or by a message), th 57 arguments or by a message), the cache will not promote
58 new writes (however, writes to 58 new writes (however, writes to already cached blocks are
59 promoted, to avoid data corrup 59 promoted, to avoid data corruption due to misordered
60 writes) and it will gradually 60 writes) and it will gradually writeback any cached
61 data. The userspace can then m 61 data. The userspace can then monitor the cleaning
62 process with "dmsetup status". 62 process with "dmsetup status". When the number of cached
63 blocks drops to zero, userspac 63 blocks drops to zero, userspace can unload the
64 dm-writecache target and repla 64 dm-writecache target and replace it with dm-linear or
65 other targets. 65 other targets.
66 max_age n 66 max_age n
67 specifies the maximum age of a 67 specifies the maximum age of a block in milliseconds. If
68 a block is stored in the cache 68 a block is stored in the cache for too long, it will be
69 written to the underlying devi 69 written to the underlying device and cleaned up.
70 metadata_only 70 metadata_only
71 only metadata is promoted to t 71 only metadata is promoted to the cache. This option
72 improves performance for heavi 72 improves performance for heavier REQ_META workloads.
73 pause_writeback n (default: 3000 73 pause_writeback n (default: 3000)
74 pause writeback if there was s 74 pause writeback if there was some write I/O redirected to
75 the origin volume in the last 75 the origin volume in the last n milliseconds
76 76
77 Status: 77 Status:
78 78
79 1. error indicator - 0 if there was no error, 79 1. error indicator - 0 if there was no error, otherwise error number
80 2. the number of blocks 80 2. the number of blocks
81 3. the number of free blocks 81 3. the number of free blocks
82 4. the number of blocks under writeback 82 4. the number of blocks under writeback
83 5. the number of read blocks 83 5. the number of read blocks
84 6. the number of read blocks that hit the cach 84 6. the number of read blocks that hit the cache
85 7. the number of write blocks 85 7. the number of write blocks
86 8. the number of write blocks that hit uncommi 86 8. the number of write blocks that hit uncommitted block
87 9. the number of write blocks that hit committ 87 9. the number of write blocks that hit committed block
88 10. the number of write blocks that bypass the 88 10. the number of write blocks that bypass the cache
89 11. the number of write blocks that are alloca 89 11. the number of write blocks that are allocated in the cache
90 12. the number of write requests that are bloc 90 12. the number of write requests that are blocked on the freelist
91 13. the number of flush requests 91 13. the number of flush requests
92 14. the number of discarded blocks 92 14. the number of discarded blocks
93 93
94 Messages: 94 Messages:
95 flush 95 flush
96 Flush the cache device. The me 96 Flush the cache device. The message returns successfully
97 if the cache device was flushe 97 if the cache device was flushed without an error
98 flush_on_suspend 98 flush_on_suspend
99 Flush the cache device on next 99 Flush the cache device on next suspend. Use this message
100 when you are going to remove t 100 when you are going to remove the cache device. The proper
101 sequence for removing the cach 101 sequence for removing the cache device is:
102 102
103 1. send the "flush_on_suspend" 103 1. send the "flush_on_suspend" message
104 2. load an inactive table with 104 2. load an inactive table with a linear target that maps
105 to the underlying device 105 to the underlying device
106 3. suspend the device 106 3. suspend the device
107 4. ask for status and verify t 107 4. ask for status and verify that there are no errors
108 5. resume the device, so that 108 5. resume the device, so that it will use the linear
109 target 109 target
110 6. the cache device is now ina 110 6. the cache device is now inactive and it can be deleted
111 cleaner 111 cleaner
112 See above "cleaner" constructo 112 See above "cleaner" constructor documentation.
113 clear_stats 113 clear_stats
114 Clear the statistics that are 114 Clear the statistics that are reported on the status line
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.