1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
|
$ crushtool --help
usage: crushtool ...
Display, modify and test a crush map
There are five stages, running one after the other:
- input/build
- tunables adjustments
- modifications
- display/test
- output
Options that are not specific to a stage.
[--infn|-i infile]
read the crush map from infile
Options for the input/build stage
--decompile|-d map decompile a crush map to source
[--outfn|-o outfile]
specify output for for (de)compilation
--compile|-c map.txt compile a map from source
--enable-unsafe-tunables
compile with unsafe tunables
--build --num_osds N layer1 ...
build a new map, where each 'layer' is
'name (uniform|straw2|straw|list|tree) size'
Options for the tunables adjustments stage
--set-choose-local-tries N
set choose local retries before re-descent
--set-choose-local-fallback-tries N
set choose local retries using fallback
permutation before re-descent
--set-choose-total-tries N
set choose total descent attempts
--set-chooseleaf-descend-once <0|1>
set chooseleaf to (not) retry the recursive descent
--set-chooseleaf-vary-r <0|1>
set chooseleaf to (not) vary r based on parent
--set-chooseleaf-stable <0|1>
set chooseleaf firstn to (not) return stable results
Options for the modifications stage
-i mapfn --add-item id weight name [--loc type name ...]
insert an item into the hierarchy at the
given location
-i mapfn --update-item id weight name [--loc type name ...]
insert or move an item into the hierarchy at the
given location
-i mapfn --remove-item name
remove the given item
-i mapfn --reweight-item name weight
reweight a given item (and adjust ancestor
weights as needed)
-i mapfn --add-bucket name type [--loc type name ...]
insert a bucket into the hierarchy at the given
location
-i mapfn --move name --loc type name ...
move the given item to specified location
-i mapfn --reweight recalculate all bucket weights
-i mapfn --rebuild-class-roots
rebuild the per-class shadow trees (normally a no-op)
-i mapfn --create-simple-rule name root type mode
create crush rule <name> to start from <root>,
replicate across buckets of type <type>, using
a choose mode of <firstn|indep>
-i mapfn --create-replicated-rule name root type
create crush rule <name> to start from <root>,
replicate across buckets of type <type>
--device-class <class>
use device class <class> for new rule
-i mapfn --remove-rule name
remove the specified crush rule
Options for the display/test stage
-f --format the format of --dump, defaults to json-pretty
can be one of json, json-pretty, xml, xml-pretty,
table, table-kv, html, html-pretty
--dump dump the crush map
--tree print map summary as a tree
--check [max_id] check if any item is referencing an unknown name/type
-i mapfn --show-location id
show location for given device id
-i mapfn --test test a range of inputs on the map
[--min-x x] [--max-x x] [--x x]
[--min-rule r] [--max-rule r] [--rule r] [--ruleset rs]
[--num-rep n]
[--pool-id n] specifies pool id
[--batches b] split the CRUSH mapping into b > 1 rounds
[--weight|-w devno weight]
where weight is 0 to 1.0
[--simulate] simulate placements using a random
number generator in place of the CRUSH
algorithm
--show-utilization show OSD usage
--show-utilization-all
include zero weight items
--show-statistics show chi squared statistics
--show-mappings show mappings
--show-bad-mappings show bad mappings
--show-choose-tries show choose tries histogram
--output-name name
prepend the data file(s) generated during the
testing routine with name
--output-csv
export select data generated during testing routine
to CSV files for off-line post-processing
use --help-output for more information
--reclassify transform legacy CRUSH map buckets and rules
by adding classes
--reclassify-bucket <bucket-match> <class> <default-parent>
--reclassify-root <bucket-name> <class>
--set-subtree-class <bucket-name> <class>
set class for all items beneath bucket-name
--compare <otherfile> compare two maps using --test parameters
Options for the output stage
[--outfn|-o outfile]
specify output for modified crush map
$ crushtool --help-output
data output from testing routine ...
absolute_weights
the decimal weight of each OSD
data layout: ROW MAJOR
OSD id (int), weight (int)
batch_device_expected_utilization_all
the expected number of objects each OSD should receive per placement batch
which may be a decimal value
data layout: COLUMN MAJOR
round (int), objects expected on OSD 0...OSD n (float)
batch_device_utilization_all
the number of objects stored on each OSD during each placement round
data layout: COLUMN MAJOR
round (int), objects stored on OSD 0...OSD n (int)
device_utilization_all
the number of objects stored on each OSD at the end of placements
data_layout: ROW MAJOR
OSD id (int), objects stored (int), objects expected (float)
device_utilization
the number of objects stored on each OSD marked 'up' at the end of placements
data_layout: ROW MAJOR
OSD id (int), objects stored (int), objects expected (float)
placement_information
the map of input -> OSD
data_layout: ROW MAJOR
input (int), OSD's mapped (int)
proportional_weights_all
the proportional weight of each OSD specified in the CRUSH map
data_layout: ROW MAJOR
OSD id (int), proportional weight (float)
proportional_weights
the proportional weight of each 'up' OSD specified in the CRUSH map
data_layout: ROW MAJOR
OSD id (int), proportional weight (float)
|
