ssd を書込み読み込みキャッシュに使いたい
HDDのアクセスを早くしたい。Writeおせーんだわ。
あれこれ、実装があるけど lvm cache / dm-cache より、bcacheの方が圧倒的に簡単で速かった。
もちろん、こんなことをしなくても、よく使うデータはHDDからメモリにキャッシュされるので、実運用ではメモリに空き容量を増やすのがとにかく先決です。bcacheは趣味ですね。
apt からインストールできる場合
sudo apt install bcache-tools
bcacheのヘッドからインストール
git から取得しインストール。
git clone https://github.com/g2p/bcache-tools.git cd bcache-tools make sudo checkinstall \ --pkgname=bcache-head \ --pkgversion="1:$(date +%Y%m%d%H%M)" \ --backup=no \ --deldoc=yes \ --fstrans=no \ --default
bcacheが有効か調べる
bcacheが使えるかどうか確認。
takuya@:~$ sudo modprobe bcache takuya@:~$ sudo lsmod | grep cache bcache 193649 0 dm_cache 41114 0 dm_persistent_data 49347 1 dm_cache dm_bio_prison 13056 1 dm_cache fscache 45542 1 nfs mbcache 17171 1 ext4 dm_mod 89405 17 dm_persistent_data,dm_cache,dm_bufio
使い始める前に用語の整理
- bcache 此のソフト
- バッキング・デバイスbacking device キャッシュされる側（通常はHDD）
- キャッシング・デバイス chacing device キャッシュ側(通常はSSD）
- make-bcache 作成コマンド
バッキングデバイスの指定
今回は、LVM上のVGから100GB のストレージを切り出してブロックデバイスを作成して/dev/sda などの代わりに使う。
これをバッキングデバイスに指定した。
takuya@:~$ sudo make-bcache -B /dev/mapper/data-storage UUID: a057b697-795d-450c-8e68-454d6f244b93 Set UUID: 9778b366-5f4c-4947-a180-c66442f2a934 version: 1 block_size: 1 data_offset: 16
キャッシュデバイスの指定
今回はSSD上のLVMから20GB ほど切り出して/dev/sdb などの代わりに使う。
これをキャッシュデバイスに指定した
takuya@:~$ sudo make-bcache -C /dev/mapper/acid-ssd UUID: 9ac55038-1b95-42fb-bd0f-3b4638476e5d Set UUID: aad94dbd-5ded-4d23-bccb-1f3c1f475c48 version: 0 nbuckets: 40960 block_size: 1 bucket_size: 1024 nr_in_set: 1 nr_this_dev: 0 first_bucket: 1
状態を確認する。
デバイスが登録されたか調べる。
bacheデバイスが見えるようになったか確認する。
takuya@:~$ ls -l -R /dev/bcache** brw-rw---- 1 root disk 253, 0 2017-03-06 18:21 /dev/bcache0 /dev/bcache: 合計 0 drwxr-xr-x 2 root root 60 2017-03-06 18:21 by-uuid /dev/bcache/by-uuid: 合計 0 lrwxrwxrwx 1 root root 13 2017-03-06 18:21 a057b697-795d-450c-8e68-454d6f244b93 -> ../../bcache0
ここで作成された /dev/bcache0/ が bcacheでキャッシュされた、バッキングデバイス(/dev/sda)などを指している。
バッキングデバイスの状態を確認する
takuya@:~$ sudo bcache-super-show /dev/mapper/data-storage sb.magic ok sb.first_sector 8 [match] sb.csum D1AA8DCE13E9B35C [match] sb.version 1 [backing device] dev.label (empty) dev.uuid a057b697-795d-450c-8e68-454d6f244b93 dev.sectors_per_block 1 dev.sectors_per_bucket 1024 dev.data.first_sector 16 dev.data.cache_mode 0 [writethrough] dev.data.cache_state 0 [detached] cset.uuid 9778b366-5f4c-4947-a180-c66442f2a934
キャッシュデバイスの状態を確認する
takuya@:~$ sudo bcache-super-show /dev/mapper/acid-ssd sb.magic ok sb.first_sector 8 [match] sb.csum 455541E08344F7F6 [match] sb.version 3 [cache device] dev.label (empty) dev.uuid 9ac55038-1b95-42fb-bd0f-3b4638476e5d dev.sectors_per_block 1 dev.sectors_per_bucket 1024 dev.cache.first_sector 1024 dev.cache.cache_sectors 41942016 dev.cache.total_sectors 41943040 dev.cache.ordered yes dev.cache.discard no dev.cache.pos 0 dev.cache.replacement 0 [lru] cset.uuid aad94dbd-5ded-4d23-bccb-1f3c1f475c48
bcache0 にキャッシュデバイスを紐付ける。
sudo bcache-super-show /dev/mapper/acid-ssd \
| grep cset.uuid | awk '{ print $2 }' \
| sudo tee /sys/block/bcache0/bcache/attach
ファイルシステムを作る
ここまでで、bcacheの準備ができたので、後はターゲット(/dev/bcache0)に対して操作をする。
takuya@:~$ sudo mkfs.ext4 /dev/bcache0
mke2fs 1.42.12 (29-Aug-2014)
Discarding device blocks: done
Creating filesystem with 26214398 4k blocks and 6553600 inodes
Filesystem UUID: 5848c161-fd7c-4031-807a-01c56dc7c967
Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,
4096000, 7962624, 11239424, 20480000, 23887872
Allocating group tables: done
Writing inode tables: done
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information: done
マウントする
sudo mount /dev/bcache0 mnt
測定してみる。
細かいテスト条件を考える暇がなかったのでかなり適当な条件ですが１GB程度の書込みで測定してみました。
単純にbw だけを見てみる。
時間がなくて、細かいテスト条件を考慮できてないので、単純にbw だけを見てみる。 読み込みは今回測定してない
ランダム書込みの測定
| デバイス | 速度(bw) |
|---|---|
| ssdのみ | 242557KB/s |
| hddのみ | 910KB/s |
| bcache/writethrough | 1314KB/s |
| bcache/writeback | 144015KB/s |
１GB程度のランダムデータは差が・・・でもまぁ少しは早くなる？
writebackで ssd にだけ書いて hdd にsync 待ちしないならそりゃ圧倒的に早いな
何かに使えそうですよね。
bcacheを使ったときの速度測定
とりあえず、fio でぱぱっと測定。測定条件を全く考慮せず使いまわした・・・次やるときはちゃんと測定条件を考えないと。
[global] bs=4k ioengine=libaio iodepth=4 size=1g direct=1 runtime=60 directory=/home/takuya/mnt filename=ssd.test.file [rand-write] rw=randwrite stonewall
fio の結果
takuya@:~$ sudo fio myjob.fio
rand-write: (g=0): rw=randwrite, bs=4K-4K/4K-4K/4K-4K, ioengine=libaio, iodepth=4
fio-2.1.11
Starting 1 process
rand-write: Laying out IO file(s) (1 file(s) / 1024MB)
Jobs: 1 (f=1): [w(1)] [100.0% done] [0KB/1576KB/0KB /s] [0/394/0 iops] [eta 00m:00s]
rand-write: (groupid=0, jobs=1): err= 0: pid=2016: Mon Mar 6 18:44:02 2017
write: io=78864KB, bw=1314.3KB/s, iops=328, runt= 60008msec
slat (usec): min=6, max=757302, avg=104.79, stdev=5956.73
clat (usec): min=78, max=2122.7K, avg=12067.09, stdev=45932.99
lat (usec): min=89, max=2122.7K, avg=12172.05, stdev=46560.28
clat percentiles (usec):
| 1.00th=[ 169], 5.00th=[ 207], 10.00th=[ 251], 20.00th=[ 5600],
| 30.00th=[ 6688], 40.00th=[ 7136], 50.00th=[ 7392], 60.00th=[ 7648],
| 70.00th=[ 7968], 80.00th=[ 8768], 90.00th=[11584], 95.00th=[20864],
| 99.00th=[130560], 99.50th=[144384], 99.90th=[602112], 99.95th=[839680],
| 99.99th=[2113536]
bw (KB /s): min= 14, max= 3093, per=100.00%, avg=1438.70, stdev=699.48
lat (usec) : 100=0.07%, 250=9.92%, 500=1.09%, 750=0.29%, 1000=0.27%
lat (msec) : 2=0.31%, 4=1.37%, 10=74.64%, 20=6.77%, 50=2.80%
lat (msec) : 100=0.22%, 250=1.94%, 500=0.12%, 750=0.14%, 1000=0.04%
lat (msec) : >=2000=0.02%
cpu : usr=0.22%, sys=1.58%, ctx=17090, majf=0, minf=8
IO depths : 1=0.1%, 2=0.1%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
issued : total=r=0/w=19716/d=0, short=r=0/w=0/d=0
latency : target=0, window=0, percentile=100.00%, depth=4
Run status group 0 (all jobs):
WRITE: io=78864KB, aggrb=1314KB/s, minb=1314KB/s, maxb=1314KB/s, mint=60008msec, maxt=60008msec
Disk stats (read/write):
bcache0: ios=0/21318, merge=0/0, ticks=0/419156, in_queue=0, util=0.00%, aggrios=0/21324, aggrmerge=0/0, aggrticks=0/260408, aggrin_queue=260408, aggrutil=98.78%
dm-4: ios=0/21434, merge=0/0, ticks=0/510812, in_queue=510812, util=98.78%, aggrios=0/19972, aggrmerge=0/1474, aggrticks=0/446100, aggrin_queue=446092, aggrutil=98.78%
sda: ios=0/19972, merge=0/1474, ticks=0/446100, in_queue=446092, util=98.78%
dm-5: ios=0/21215, merge=0/0, ticks=0/10004, in_queue=10004, util=5.14%, aggrios=0/26443, aggrmerge=0/2553, aggrticks=0/79524, aggrin_queue=79516, aggrutil=7.90%
sdb: ios=0/26443, merge=0/2553, ticks=0/79524, in_queue=79516, util=7.90%
SSDだけの場合
キャッシュデバイスに利用したSSDからLVMを切り出して使うと次の速度だった。
takuya@:~$ fio myjob.fio
rand-write: (g=0): rw=randwrite, bs=4K-4K/4K-4K/4K-4K, ioengine=libaio, iodepth=4
fio-2.1.11
Starting 1 process
Jobs: 1 (f=1): [w(1)] [100.0% done] [0KB/237.4MB/0KB /s] [0/60.8K/0 iops] [eta 00m:00s]
rand-write: (groupid=0, jobs=1): err= 0: pid=2146: Mon Mar 6 18:47:09 2017
write: io=1024.0MB, bw=242557KB/s, iops=60639, runt= 4323msec
slat (usec): min=2, max=1436, avg= 7.19, stdev=13.71
clat (usec): min=12, max=48210, avg=57.96, stdev=215.17
lat (usec): min=26, max=48214, avg=65.23, stdev=216.07
clat percentiles (usec):
| 1.00th=[ 33], 5.00th=[ 36], 10.00th=[ 37], 20.00th=[ 39],
| 30.00th=[ 40], 40.00th=[ 41], 50.00th=[ 41], 60.00th=[ 42],
| 70.00th=[ 43], 80.00th=[ 46], 90.00th=[ 58], 95.00th=[ 231],
| 99.00th=[ 237], 99.50th=[ 241], 99.90th=[ 684], 99.95th=[ 1160],
| 99.99th=[ 5856]
bw (KB /s): min=169536, max=258520, per=99.67%, avg=241758.00, stdev=30286.36
lat (usec) : 20=0.01%, 50=83.98%, 100=9.08%, 250=6.55%, 500=0.24%
lat (usec) : 750=0.05%, 1000=0.02%
lat (msec) : 2=0.06%, 4=0.01%, 10=0.01%, 20=0.01%, 50=0.01%
cpu : usr=11.29%, sys=44.42%, ctx=154698, majf=0, minf=8
IO depths : 1=0.1%, 2=0.1%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
issued : total=r=0/w=262144/d=0, short=r=0/w=0/d=0
latency : target=0, window=0, percentile=100.00%, depth=4
Run status group 0 (all jobs):
WRITE: io=1024.0MB, aggrb=242557KB/s, minb=242557KB/s, maxb=242557KB/s, mint=4323msec, maxt=4323msec
Disk stats (read/write):
dm-1: ios=644/251352, merge=0/0, ticks=144/14056, in_queue=14208, util=97.55%, aggrios=644/262180, aggrmerge=0/15, aggrticks=144/18640, aggrin_queue=18752, aggrutil=96.11%
sdb: ios=644/262180, merge=0/15, ticks=144/18640, in_queue=18752, util=96.11%
HDDだけの場合
takuya@:~$ sudo fio myjob.fio
rand-write: (g=0): rw=randwrite, bs=4K-4K/4K-4K/4K-4K, ioengine=libaio, iodepth=4
fio-2.1.11
Starting 1 process
rand-write: Laying out IO file(s) (1 file(s) / 1024MB)
Jobs: 1 (f=1): [w(1)] [100.0% done] [0KB/1420KB/0KB /s] [0/355/0 iops] [eta 00m:00s]
rand-write: (groupid=0, jobs=1): err= 0: pid=2249: Mon Mar 6 18:50:09 2017
write: io=54776KB, bw=931985B/s, iops=227, runt= 60184msec
slat (usec): min=5, max=371161, avg=137.76, stdev=5134.81
clat (usec): min=101, max=883763, avg=17428.75, stdev=51339.16
lat (usec): min=126, max=883769, avg=17566.73, stdev=51900.52
clat percentiles (usec):
| 1.00th=[ 171], 5.00th=[ 209], 10.00th=[ 223], 20.00th=[ 5600],
| 30.00th=[ 6752], 40.00th=[ 7200], 50.00th=[ 7456], 60.00th=[ 7712],
| 70.00th=[ 8160], 80.00th=[ 9280], 90.00th=[17792], 95.00th=[125440],
| 99.00th=[189440], 99.50th=[261120], 99.90th=[684032], 99.95th=[880640],
| 99.99th=[880640]
bw (KB /s): min= 17, max= 3032, per=100.00%, avg=991.52, stdev=724.89
lat (usec) : 250=11.17%, 500=0.91%, 750=0.04%, 1000=0.01%
lat (msec) : 2=0.24%, 4=1.22%, 10=71.24%, 20=7.80%, 50=1.29%
lat (msec) : 100=0.06%, 250=5.26%, 500=0.48%, 750=0.19%, 1000=0.09%
cpu : usr=0.19%, sys=1.14%, ctx=11227, majf=0, minf=8
IO depths : 1=0.1%, 2=0.1%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
issued : total=r=0/w=13694/d=0, short=r=0/w=0/d=0
latency : target=0, window=0, percentile=100.00%, depth=4
Run status group 0 (all jobs):
WRITE: io=54776KB, aggrb=910KB/s, minb=910KB/s, maxb=910KB/s, mint=60184msec, maxt=60184msec
Disk stats (read/write):
dm-6: ios=0/14883, merge=0/0, ticks=0/462800, in_queue=473700, util=99.95%, aggrios=0/13821, aggrmerge=0/1075, aggrticks=0/443268, aggrin_queue=450992, aggrutil=99.95%
sda: ios=0/13821, merge=0/1075, ticks=0/443268, in_queue=450992, util=99.95%
takuya@:~$
writeback にして測定してみる。
witebackを有効にする
takuya@:~$ echo writeback | sudo tee /sys/block/bcache0/bcache/cache_mode writeback
takuya@:~$ sudo fio myjob.fio
rand-write: (g=0): rw=randwrite, bs=4K-4K/4K-4K/4K-4K, ioengine=libaio, iodepth=4
fio-2.1.11
Starting 1 process
Jobs: 1 (f=1): [w(1)] [100.0% done] [0KB/151.5MB/0KB /s] [0/38.8K/0 iops] [eta 00m:00s]
rand-write: (groupid=0, jobs=1): err= 0: pid=2341: Mon Mar 6 18:54:19 2017
write: io=1024.0MB, bw=144015KB/s, iops=36003, runt= 7281msec
slat (usec): min=1, max=6871, avg=12.32, stdev=16.83
clat (usec): min=19, max=14755, avg=98.10, stdev=368.31
lat (usec): min=28, max=14760, avg=110.48, stdev=368.54
clat percentiles (usec):
| 1.00th=[ 34], 5.00th=[ 38], 10.00th=[ 39], 20.00th=[ 40],
| 30.00th=[ 41], 40.00th=[ 43], 50.00th=[ 43], 60.00th=[ 44],
| 70.00th=[ 46], 80.00th=[ 50], 90.00th=[ 63], 95.00th=[ 157],
| 99.00th=[ 1864], 99.50th=[ 2480], 99.90th=[ 3152], 99.95th=[ 5024],
| 99.99th=[13504]
bw (KB /s): min=136680, max=161360, per=99.65%, avg=143503.93, stdev=6617.05
lat (usec) : 20=0.01%, 50=79.82%, 100=14.01%, 250=2.88%, 500=0.19%
lat (usec) : 750=0.67%, 1000=0.42%
lat (msec) : 2=1.12%, 4=0.81%, 10=0.06%, 20=0.03%
cpu : usr=5.00%, sys=44.67%, ctx=67203, majf=0, minf=7
IO depths : 1=0.1%, 2=0.1%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
issued : total=r=0/w=262144/d=0, short=r=0/w=0/d=0
latency : target=0, window=0, percentile=100.00%, depth=4
Run status group 0 (all jobs):
WRITE: io=1024.0MB, aggrb=144015KB/s, minb=144015KB/s, maxb=144015KB/s, mint=7281msec, maxt=7281msec
Disk stats (read/write):
bcache0: ios=223/257488, merge=0/0, ticks=28/26512, in_queue=0, util=0.00%, aggrios=111/132361, aggrmerge=0/0, aggrticks=14/13326, aggrin_queue=13352, aggrutil=95.97%
dm-4: ios=0/1, merge=0/0, ticks=0/0, in_queue=0, util=0.00%, aggrios=0/1, aggrmerge=0/0, aggrticks=0/0, aggrin_queue=0, aggrutil=0.00%
sda: ios=0/1, merge=0/0, ticks=0/0, in_queue=0, util=0.00%
dm-5: ios=223/264721, merge=0/0, ticks=28/26652, in_queue=26704, util=95.97%, aggrios=223/258582, aggrmerge=0/6444, aggrticks=28/15168, aggrin_queue=15116, aggrutil=95.91%
sdb: ios=223/258582, merge=0/6444, ticks=28/15168, in_queue=15116, util=95.91%
takuya@:~$
writethrough で限界までパフォーマンス上げてみる
シーケンシャルIOの検出・バイパスオフ
takuya@:~$ echo writethrough | sudo tee /sys/block/bcache0/bcache/cache_mode writethrough takuya@:~$ echo 0 | sudo tee /sys/block/bcache0/bcache/sequential_cutoff 0 takuya@:~$ echo 0 | sudo tee /sys/fs/ bcache/ btrfs/ cgroup/ ext4/ fuse/ pstore/
レイテンシ検出オフ
takuya@:~$ echo 0 | sudo tee /sys/fs/bcache/aad94dbd-5ded-4d23-bccb-1f3c1f475c48/congested_read_threshold_us 0 takuya@:~$ echo 0 | sudo tee /sys/fs/bcache/aad94dbd-5ded-4d23-bccb-1f3c1f475c48/congested_write_threshold_us 0
この条件だと、測定結果は次のとおりになった。5%くらいはやい？あんまり変わらないね。
takuya@:~$ sudo fio myjob.fio
rand-write: (g=0): rw=randwrite, bs=4K-4K/4K-4K/4K-4K, ioengine=libaio, iodepth=4
fio-2.1.11
Starting 1 process
Jobs: 1 (f=0): [w(1)] [8.1% done] [0KB/1810KB/0KB /s] [0/452/0 iops] [eta 11m:35s]
rand-write: (groupid=0, jobs=1): err= 0: pid=4873: Mon Mar 6 19:12:02 2017
write: io=84528KB, bw=1408.7KB/s, iops=352, runt= 60007msec
slat (usec): min=5, max=16313, avg=86.45, stdev=167.83
clat (usec): min=77, max=726096, avg=11262.20, stdev=31084.99
lat (usec): min=85, max=730582, avg=11349.47, stdev=31109.02
clat percentiles (usec):
| 1.00th=[ 163], 5.00th=[ 209], 10.00th=[ 4576], 20.00th=[ 5792],
| 30.00th=[ 6816], 40.00th=[ 7136], 50.00th=[ 7392], 60.00th=[ 7648],
| 70.00th=[ 7968], 80.00th=[ 8768], 90.00th=[10944], 95.00th=[18816],
| 99.00th=[129536], 99.50th=[140288], 99.90th=[544768], 99.95th=[634880],
| 99.99th=[724992]
bw (KB /s): min= 71, max= 3040, per=100.00%, avg=1479.37, stdev=684.76
lat (usec) : 100=0.06%, 250=6.80%, 500=0.60%, 750=0.18%, 1000=0.13%
lat (msec) : 2=0.11%, 4=1.35%, 10=79.16%, 20=7.32%, 50=2.14%
lat (msec) : 100=0.01%, 250=1.91%, 500=0.11%, 750=0.13%
cpu : usr=0.73%, sys=3.81%, ctx=17268, majf=0, minf=8
IO depths : 1=0.1%, 2=0.1%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
issued : total=r=0/w=21132/d=0, short=r=0/w=0/d=0
latency : target=0, window=0, percentile=100.00%, depth=4
Run status group 0 (all jobs):
WRITE: io=84528KB, aggrb=1408KB/s, minb=1408KB/s, maxb=1408KB/s, mint=60007msec, maxt=60007msec
Disk stats (read/write):
bcache0: ios=0/21070, merge=0/0, ticks=0/242964, in_queue=0, util=0.00%, aggrios=30/21496, aggrmerge=0/0, aggrticks=0/178762, aggrin_queue=178812, aggrutil=99.98%
dm-4: ios=0/21224, merge=0/0, ticks=0/355520, in_queue=355620, util=99.98%, aggrios=0/21184, aggrmerge=0/51, aggrticks=0/296676, aggrin_queue=296624, aggrutil=99.97%
sda: ios=0/21184, merge=0/51, ticks=0/296676, in_queue=296624, util=99.97%
dm-5: ios=60/21768, merge=0/0, ticks=0/2004, in_queue=2004, util=2.88%, aggrios=60/22260, aggrmerge=0/71, aggrticks=0/1788, aggrin_queue=1752, aggrutil=2.54%
sdb: ios=60/22260, merge=0/71, ticks=0/1788, in_queue=1752, util=2.54%
停止する
stop に 1 を書き込んだらデバイスを停止できる。
takuya@:~$ echo 1 | sudo tee /sys/block/bcache0/bcache/stop 1
もしまた使いたいときは、wipefs でフラッシュする
takuya@:~$ sudo wipefs -a /dev/mapper/acid-cache
主な操作方法
主な操作は /sys/fs 経由で行う。
参考資料
http://www.slideshare.net/nobuto_m/bcachessd-hdd
http://unix.stackexchange.com/questions/225017/how-to-remove-bcache0-volume
http://www.tech-g.com/2015/05/10/bcache-linux-ssd-caching-for-hard-drives-on-debian-jessie/
https://pommi.nethuis.nl/ssd-caching-using-linux-and-bcache/
https://jp.linux.com/news/linuxcom-exclusive/413722-lco20140225
