1 =================== 1 ===================
2 Block IO Controller 2 Block IO Controller
3 =================== 3 ===================
4 4
5 Overview 5 Overview
6 ======== 6 ========
7 cgroup subsys "blkio" implements the block io 7 cgroup subsys "blkio" implements the block io controller. There seems to be
8 a need of various kinds of IO control policies 8 a need of various kinds of IO control policies (like proportional BW, max BW)
9 both at leaf nodes as well as at intermediate 9 both at leaf nodes as well as at intermediate nodes in a storage hierarchy.
10 Plan is to use the same cgroup based managemen 10 Plan is to use the same cgroup based management interface for blkio controller
11 and based on user options switch IO policies i 11 and based on user options switch IO policies in the background.
12 12
13 One IO control policy is throttling policy whi 13 One IO control policy is throttling policy which can be used to
14 specify upper IO rate limits on devices. This 14 specify upper IO rate limits on devices. This policy is implemented in
15 generic block layer and can be used on leaf no 15 generic block layer and can be used on leaf nodes as well as higher
16 level logical devices like device mapper. 16 level logical devices like device mapper.
17 17
18 HOWTO 18 HOWTO
19 ===== 19 =====
20 20
21 Throttling/Upper Limit policy 21 Throttling/Upper Limit policy
22 ----------------------------- 22 -----------------------------
23 Enable Block IO controller:: 23 Enable Block IO controller::
24 24
25 CONFIG_BLK_CGROUP=y 25 CONFIG_BLK_CGROUP=y
26 26
27 Enable throttling in block layer:: 27 Enable throttling in block layer::
28 28
29 CONFIG_BLK_DEV_THROTTLING=y 29 CONFIG_BLK_DEV_THROTTLING=y
30 30
31 Mount blkio controller (see cgroups.txt, Why a 31 Mount blkio controller (see cgroups.txt, Why are cgroups needed?)::
32 32
33 mount -t cgroup -o blkio none /sys/fs/ 33 mount -t cgroup -o blkio none /sys/fs/cgroup/blkio
34 34
35 Specify a bandwidth rate on particular device 35 Specify a bandwidth rate on particular device for root group. The format
36 for policy is "<major>:<minor> <bytes_per_sec 36 for policy is "<major>:<minor> <bytes_per_second>"::
37 37
38 echo "8:16 1048576" > /sys/fs/cgroup/ 38 echo "8:16 1048576" > /sys/fs/cgroup/blkio/blkio.throttle.read_bps_device
39 39
40 This will put a limit of 1MB/second on reads h 40 This will put a limit of 1MB/second on reads happening for root group
41 on device having major/minor number 8:16. 41 on device having major/minor number 8:16.
42 42
43 Run dd to read a file and see if rate is throt 43 Run dd to read a file and see if rate is throttled to 1MB/s or not::
44 44
45 # dd iflag=direct if=/mnt/common/zerof 45 # dd iflag=direct if=/mnt/common/zerofile of=/dev/null bs=4K count=1024
46 1024+0 records in 46 1024+0 records in
47 1024+0 records out 47 1024+0 records out
48 4194304 bytes (4.2 MB) copied, 4.0001 48 4194304 bytes (4.2 MB) copied, 4.0001 s, 1.0 MB/s
49 49
50 Limits for writes can be put using blkio.throt 50 Limits for writes can be put using blkio.throttle.write_bps_device file.
51 51
52 Hierarchical Cgroups 52 Hierarchical Cgroups
53 ==================== 53 ====================
54 54
55 Throttling implements hierarchy support; howev 55 Throttling implements hierarchy support; however,
56 throttling's hierarchy support is enabled iff 56 throttling's hierarchy support is enabled iff "sane_behavior" is
57 enabled from cgroup side, which currently is a 57 enabled from cgroup side, which currently is a development option and
58 not publicly available. 58 not publicly available.
59 59
60 If somebody created a hierarchy like as follow 60 If somebody created a hierarchy like as follows::
61 61
62 root 62 root
63 / \ 63 / \
64 test1 test2 64 test1 test2
65 | 65 |
66 test3 66 test3
67 67
68 Throttling with "sane_behavior" will handle th 68 Throttling with "sane_behavior" will handle the
69 hierarchy correctly. For throttling, all limit 69 hierarchy correctly. For throttling, all limits apply
70 to the whole subtree while all statistics are 70 to the whole subtree while all statistics are local to the IOs
71 directly generated by tasks in that cgroup. 71 directly generated by tasks in that cgroup.
72 72
73 Throttling without "sane_behavior" enabled fro 73 Throttling without "sane_behavior" enabled from cgroup side will
74 practically treat all groups at same level as 74 practically treat all groups at same level as if it looks like the
75 following:: 75 following::
76 76
77 pivot 77 pivot
78 / / \ \ 78 / / \ \
79 root test1 test2 tes 79 root test1 test2 test3
80 80
81 Various user visible config options 81 Various user visible config options
82 =================================== 82 ===================================
83 83
84 CONFIG_BLK_CGROUP 84 CONFIG_BLK_CGROUP
85 Block IO controller. 85 Block IO controller.
86 86
87 CONFIG_BFQ_CGROUP_DEBUG 87 CONFIG_BFQ_CGROUP_DEBUG
88 Debug help. Right now some additiona 88 Debug help. Right now some additional stats file show up in cgroup
89 if this option is enabled. 89 if this option is enabled.
90 90
91 CONFIG_BLK_DEV_THROTTLING 91 CONFIG_BLK_DEV_THROTTLING
92 Enable block device throttling suppo 92 Enable block device throttling support in block layer.
93 93
94 Details of cgroup files 94 Details of cgroup files
95 ======================= 95 =======================
96 96
97 Proportional weight policy files 97 Proportional weight policy files
98 -------------------------------- 98 --------------------------------
99 99
100 blkio.bfq.weight 100 blkio.bfq.weight
101 Specifies per cgroup weight. This is 101 Specifies per cgroup weight. This is default weight of the group
102 on all the devices until and unless 102 on all the devices until and unless overridden by per device rule
103 (see `blkio.bfq.weight_device` below 103 (see `blkio.bfq.weight_device` below).
104 104
105 Currently allowed range of weights i 105 Currently allowed range of weights is from 1 to 1000. For more details,
106 see Documentation/block/bfq-iosched. 106 see Documentation/block/bfq-iosched.rst.
107 107
108 blkio.bfq.weight_device 108 blkio.bfq.weight_device
109 Specifies per cgroup per device weig 109 Specifies per cgroup per device weights, overriding the default group
110 weight. For more details, see Docume 110 weight. For more details, see Documentation/block/bfq-iosched.rst.
111 111
112 Following is the format:: 112 Following is the format::
113 113
114 # echo dev_maj:dev_minor weight > 114 # echo dev_maj:dev_minor weight > blkio.bfq.weight_device
115 115
116 Configure weight=300 on /dev/sdb (8: 116 Configure weight=300 on /dev/sdb (8:16) in this cgroup::
117 117
118 # echo 8:16 300 > blkio.bfq.weight 118 # echo 8:16 300 > blkio.bfq.weight_device
119 # cat blkio.bfq.weight_device 119 # cat blkio.bfq.weight_device
120 dev weight 120 dev weight
121 8:16 300 121 8:16 300
122 122
123 Configure weight=500 on /dev/sda (8: 123 Configure weight=500 on /dev/sda (8:0) in this cgroup::
124 124
125 # echo 8:0 500 > blkio.bfq.weight_ 125 # echo 8:0 500 > blkio.bfq.weight_device
126 # cat blkio.bfq.weight_device 126 # cat blkio.bfq.weight_device
127 dev weight 127 dev weight
128 8:0 500 128 8:0 500
129 8:16 300 129 8:16 300
130 130
131 Remove specific weight for /dev/sda 131 Remove specific weight for /dev/sda in this cgroup::
132 132
133 # echo 8:0 0 > blkio.bfq.weight_de 133 # echo 8:0 0 > blkio.bfq.weight_device
134 # cat blkio.bfq.weight_device 134 # cat blkio.bfq.weight_device
135 dev weight 135 dev weight
136 8:16 300 136 8:16 300
137 137
138 blkio.time 138 blkio.time
139 Disk time allocated to cgroup per de 139 Disk time allocated to cgroup per device in milliseconds. First
140 two fields specify the major and min 140 two fields specify the major and minor number of the device and
141 third field specifies the disk time 141 third field specifies the disk time allocated to group in
142 milliseconds. 142 milliseconds.
143 143
144 blkio.sectors 144 blkio.sectors
145 Number of sectors transferred to/fro 145 Number of sectors transferred to/from disk by the group. First
146 two fields specify the major and min 146 two fields specify the major and minor number of the device and
147 third field specifies the number of 147 third field specifies the number of sectors transferred by the
148 group to/from the device. 148 group to/from the device.
149 149
150 blkio.io_service_bytes 150 blkio.io_service_bytes
151 Number of bytes transferred to/from 151 Number of bytes transferred to/from the disk by the group. These
152 are further divided by the type of o 152 are further divided by the type of operation - read or write, sync
153 or async. First two fields specify t 153 or async. First two fields specify the major and minor number of the
154 device, third field specifies the op 154 device, third field specifies the operation type and the fourth field
155 specifies the number of bytes. 155 specifies the number of bytes.
156 156
157 blkio.io_serviced 157 blkio.io_serviced
158 Number of IOs (bio) issued to the di 158 Number of IOs (bio) issued to the disk by the group. These
159 are further divided by the type of o 159 are further divided by the type of operation - read or write, sync
160 or async. First two fields specify t 160 or async. First two fields specify the major and minor number of the
161 device, third field specifies the op 161 device, third field specifies the operation type and the fourth field
162 specifies the number of IOs. 162 specifies the number of IOs.
163 163
164 blkio.io_service_time 164 blkio.io_service_time
165 Total amount of time between request 165 Total amount of time between request dispatch and request completion
166 for the IOs done by this cgroup. Thi 166 for the IOs done by this cgroup. This is in nanoseconds to make it
167 meaningful for flash devices too. Fo 167 meaningful for flash devices too. For devices with queue depth of 1,
168 this time represents the actual serv 168 this time represents the actual service time. When queue_depth > 1,
169 that is no longer true as requests m 169 that is no longer true as requests may be served out of order. This
170 may cause the service time for a giv 170 may cause the service time for a given IO to include the service time
171 of multiple IOs when served out of o 171 of multiple IOs when served out of order which may result in total
172 io_service_time > actual time elapse 172 io_service_time > actual time elapsed. This time is further divided by
173 the type of operation - read or writ 173 the type of operation - read or write, sync or async. First two fields
174 specify the major and minor number o 174 specify the major and minor number of the device, third field
175 specifies the operation type and the 175 specifies the operation type and the fourth field specifies the
176 io_service_time in ns. 176 io_service_time in ns.
177 177
178 blkio.io_wait_time 178 blkio.io_wait_time
179 Total amount of time the IOs for thi 179 Total amount of time the IOs for this cgroup spent waiting in the
180 scheduler queues for service. This c 180 scheduler queues for service. This can be greater than the total time
181 elapsed since it is cumulative io_wa 181 elapsed since it is cumulative io_wait_time for all IOs. It is not a
182 measure of total time the cgroup spe 182 measure of total time the cgroup spent waiting but rather a measure of
183 the wait_time for its individual IOs 183 the wait_time for its individual IOs. For devices with queue_depth > 1
184 this metric does not include the tim 184 this metric does not include the time spent waiting for service once
185 the IO is dispatched to the device b 185 the IO is dispatched to the device but till it actually gets serviced
186 (there might be a time lag here due 186 (there might be a time lag here due to re-ordering of requests by the
187 device). This is in nanoseconds to m 187 device). This is in nanoseconds to make it meaningful for flash
188 devices too. This time is further di 188 devices too. This time is further divided by the type of operation -
189 read or write, sync or async. First 189 read or write, sync or async. First two fields specify the major and
190 minor number of the device, third fi 190 minor number of the device, third field specifies the operation type
191 and the fourth field specifies the i 191 and the fourth field specifies the io_wait_time in ns.
192 192
193 blkio.io_merged 193 blkio.io_merged
194 Total number of bios/requests merged 194 Total number of bios/requests merged into requests belonging to this
195 cgroup. This is further divided by t 195 cgroup. This is further divided by the type of operation - read or
196 write, sync or async. 196 write, sync or async.
197 197
198 blkio.io_queued 198 blkio.io_queued
199 Total number of requests queued up a 199 Total number of requests queued up at any given instant for this
200 cgroup. This is further divided by t 200 cgroup. This is further divided by the type of operation - read or
201 write, sync or async. 201 write, sync or async.
202 202
203 blkio.avg_queue_size 203 blkio.avg_queue_size
204 Debugging aid only enabled if CONFIG 204 Debugging aid only enabled if CONFIG_BFQ_CGROUP_DEBUG=y.
205 The average queue size for this cgro 205 The average queue size for this cgroup over the entire time of this
206 cgroup's existence. Queue size sampl 206 cgroup's existence. Queue size samples are taken each time one of the
207 queues of this cgroup gets a timesli 207 queues of this cgroup gets a timeslice.
208 208
209 blkio.group_wait_time 209 blkio.group_wait_time
210 Debugging aid only enabled if CONFIG 210 Debugging aid only enabled if CONFIG_BFQ_CGROUP_DEBUG=y.
211 This is the amount of time the cgrou 211 This is the amount of time the cgroup had to wait since it became busy
212 (i.e., went from 0 to 1 request queu 212 (i.e., went from 0 to 1 request queued) to get a timeslice for one of
213 its queues. This is different from t 213 its queues. This is different from the io_wait_time which is the
214 cumulative total of the amount of ti 214 cumulative total of the amount of time spent by each IO in that cgroup
215 waiting in the scheduler queue. This 215 waiting in the scheduler queue. This is in nanoseconds. If this is
216 read when the cgroup is in a waiting 216 read when the cgroup is in a waiting (for timeslice) state, the stat
217 will only report the group_wait_time 217 will only report the group_wait_time accumulated till the last time it
218 got a timeslice and will not include 218 got a timeslice and will not include the current delta.
219 219
220 blkio.empty_time 220 blkio.empty_time
221 Debugging aid only enabled if CONFIG 221 Debugging aid only enabled if CONFIG_BFQ_CGROUP_DEBUG=y.
222 This is the amount of time a cgroup 222 This is the amount of time a cgroup spends without any pending
223 requests when not being served, i.e. 223 requests when not being served, i.e., it does not include any time
224 spent idling for one of the queues o 224 spent idling for one of the queues of the cgroup. This is in
225 nanoseconds. If this is read when th 225 nanoseconds. If this is read when the cgroup is in an empty state,
226 the stat will only report the empty_ 226 the stat will only report the empty_time accumulated till the last
227 time it had a pending request and wi 227 time it had a pending request and will not include the current delta.
228 228
229 blkio.idle_time 229 blkio.idle_time
230 Debugging aid only enabled if CONFIG 230 Debugging aid only enabled if CONFIG_BFQ_CGROUP_DEBUG=y.
231 This is the amount of time spent by 231 This is the amount of time spent by the IO scheduler idling for a
232 given cgroup in anticipation of a be 232 given cgroup in anticipation of a better request than the existing ones
233 from other queues/cgroups. This is i 233 from other queues/cgroups. This is in nanoseconds. If this is read
234 when the cgroup is in an idling stat 234 when the cgroup is in an idling state, the stat will only report the
235 idle_time accumulated till the last 235 idle_time accumulated till the last idle period and will not include
236 the current delta. 236 the current delta.
237 237
238 blkio.dequeue 238 blkio.dequeue
239 Debugging aid only enabled if CONFIG 239 Debugging aid only enabled if CONFIG_BFQ_CGROUP_DEBUG=y. This
240 gives the statistics about how many 240 gives the statistics about how many a times a group was dequeued
241 from service tree of the device. Fir 241 from service tree of the device. First two fields specify the major
242 and minor number of the device and t 242 and minor number of the device and third field specifies the number
243 of times a group was dequeued from a 243 of times a group was dequeued from a particular device.
244 244
245 blkio.*_recursive 245 blkio.*_recursive
246 Recursive version of various stats. 246 Recursive version of various stats. These files show the
247 same information as their non-recurs 247 same information as their non-recursive counterparts but
248 include stats from all the descendan 248 include stats from all the descendant cgroups.
249 249
250 Throttling/Upper limit policy files 250 Throttling/Upper limit policy files
251 ----------------------------------- 251 -----------------------------------
252 blkio.throttle.read_bps_device 252 blkio.throttle.read_bps_device
253 Specifies upper limit on READ rate f 253 Specifies upper limit on READ rate from the device. IO rate is
254 specified in bytes per second. Rules 254 specified in bytes per second. Rules are per device. Following is
255 the format:: 255 the format::
256 256
257 echo "<major>:<minor> <rate_bytes 257 echo "<major>:<minor> <rate_bytes_per_second>" > /cgrp/blkio.throttle.read_bps_device
258 258
259 blkio.throttle.write_bps_device 259 blkio.throttle.write_bps_device
260 Specifies upper limit on WRITE rate 260 Specifies upper limit on WRITE rate to the device. IO rate is
261 specified in bytes per second. Rules 261 specified in bytes per second. Rules are per device. Following is
262 the format:: 262 the format::
263 263
264 echo "<major>:<minor> <rate_bytes 264 echo "<major>:<minor> <rate_bytes_per_second>" > /cgrp/blkio.throttle.write_bps_device
265 265
266 blkio.throttle.read_iops_device 266 blkio.throttle.read_iops_device
267 Specifies upper limit on READ rate f 267 Specifies upper limit on READ rate from the device. IO rate is
268 specified in IO per second. Rules ar 268 specified in IO per second. Rules are per device. Following is
269 the format:: 269 the format::
270 270
271 echo "<major>:<minor> <rate_io_per 271 echo "<major>:<minor> <rate_io_per_second>" > /cgrp/blkio.throttle.read_iops_device
272 272
273 blkio.throttle.write_iops_device 273 blkio.throttle.write_iops_device
274 Specifies upper limit on WRITE rate 274 Specifies upper limit on WRITE rate to the device. IO rate is
275 specified in io per second. Rules ar 275 specified in io per second. Rules are per device. Following is
276 the format:: 276 the format::
277 277
278 echo "<major>:<minor> <rate_io_pe 278 echo "<major>:<minor> <rate_io_per_second>" > /cgrp/blkio.throttle.write_iops_device
279 279
280 Note: If both BW and IOPS rules are 280 Note: If both BW and IOPS rules are specified for a device, then IO is
281 subjected to both the constraints. 281 subjected to both the constraints.
282 282
283 blkio.throttle.io_serviced 283 blkio.throttle.io_serviced
284 Number of IOs (bio) issued to the di 284 Number of IOs (bio) issued to the disk by the group. These
285 are further divided by the type of o 285 are further divided by the type of operation - read or write, sync
286 or async. First two fields specify t 286 or async. First two fields specify the major and minor number of the
287 device, third field specifies the op 287 device, third field specifies the operation type and the fourth field
288 specifies the number of IOs. 288 specifies the number of IOs.
289 289
290 blkio.throttle.io_service_bytes 290 blkio.throttle.io_service_bytes
291 Number of bytes transferred to/from 291 Number of bytes transferred to/from the disk by the group. These
292 are further divided by the type of o 292 are further divided by the type of operation - read or write, sync
293 or async. First two fields specify t 293 or async. First two fields specify the major and minor number of the
294 device, third field specifies the op 294 device, third field specifies the operation type and the fourth field
295 specifies the number of bytes. 295 specifies the number of bytes.
296 296
297 Common files among various policies 297 Common files among various policies
298 ----------------------------------- 298 -----------------------------------
299 blkio.reset_stats 299 blkio.reset_stats
300 Writing an int to this file will res 300 Writing an int to this file will result in resetting all the stats
301 for that cgroup. 301 for that cgroup.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.