|
1
|
|
|
# Licensed to the StackStorm, Inc ('StackStorm') under one or more |
|
2
|
|
|
# contributor license agreements. See the NOTICE file distributed with |
|
3
|
|
|
# this work for additional information regarding copyright ownership. |
|
4
|
|
|
# The ASF licenses this file to You under the Apache License, Version 2.0 |
|
5
|
|
|
# (the "License"); you may not use this file except in compliance with |
|
6
|
|
|
# the License. You may obtain a copy of the License at |
|
7
|
|
|
# |
|
8
|
|
|
# http://www.apache.org/licenses/LICENSE-2.0 |
|
9
|
|
|
# |
|
10
|
|
|
# Unless required by applicable law or agreed to in writing, software |
|
11
|
|
|
# distributed under the License is distributed on an "AS IS" BASIS, |
|
12
|
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
|
13
|
|
|
# See the License for the specific language governing permissions and |
|
14
|
|
|
# limitations under the License. |
|
15
|
|
|
|
|
16
|
|
|
import ctypes |
|
17
|
|
|
import hashlib |
|
18
|
|
|
|
|
19
|
|
|
from st2reactor.container.partitioners import DefaultPartitioner, get_all_enabled_sensors |
|
20
|
|
|
|
|
21
|
|
|
__all__ = [ |
|
22
|
|
|
'HashPartitioner', |
|
23
|
|
|
'Range' |
|
24
|
|
|
] |
|
25
|
|
|
|
|
26
|
|
|
# The range expression serialized is of the form `RANGE_START..RANGE_END|RANGE_START..RANGE_END ...` |
|
27
|
|
|
SUB_RANGE_SEPARATOR = '|' |
|
28
|
|
|
RANGE_BOUNDARY_SEPARATOR = '..' |
|
29
|
|
|
|
|
30
|
|
|
|
|
31
|
|
|
class Range(object): |
|
32
|
|
|
|
|
33
|
|
|
RANGE_MIN_ENUM = 'min' |
|
34
|
|
|
RANGE_MIN_VALUE = 0 |
|
35
|
|
|
|
|
36
|
|
|
RANGE_MAX_ENUM = 'max' |
|
37
|
|
|
RANGE_MAX_VALUE = 2**32 |
|
38
|
|
|
|
|
39
|
|
|
def __init__(self, range_repr): |
|
40
|
|
|
self.range_start, self.range_end = self._get_range_boundaries(range_repr) |
|
41
|
|
|
|
|
42
|
|
|
def __contains__(self, item): |
|
43
|
|
|
return item >= self.range_start and item < self.range_end |
|
44
|
|
|
|
|
45
|
|
|
def _get_range_boundaries(self, range_repr): |
|
46
|
|
|
range_repr = [value.strip() for value in range_repr.split(RANGE_BOUNDARY_SEPARATOR)] |
|
47
|
|
|
if len(range_repr) != 2: |
|
48
|
|
|
raise ValueError('Unsupported sub-range format %s.' % range_repr) |
|
49
|
|
|
|
|
50
|
|
|
range_start = self._get_valid_range_boundary(range_repr[0]) |
|
51
|
|
|
range_end = self._get_valid_range_boundary(range_repr[1]) |
|
52
|
|
|
|
|
53
|
|
|
if range_start > range_end: |
|
54
|
|
|
raise ValueError('Misconfigured range [%d..%d]' % (range_start, range_end)) |
|
55
|
|
|
return (range_start, range_end) |
|
56
|
|
|
|
|
57
|
|
|
def _get_valid_range_boundary(self, boundary_value): |
|
58
|
|
|
# Not elegant by any means but super clear. |
|
59
|
|
|
if boundary_value.lower() == self.RANGE_MIN_ENUM: |
|
60
|
|
|
return self.RANGE_MIN_VALUE |
|
61
|
|
|
if boundary_value.lower() == self.RANGE_MAX_ENUM: |
|
62
|
|
|
return self.RANGE_MAX_VALUE |
|
63
|
|
|
boundary_value = int(boundary_value) |
|
64
|
|
|
# Disallow any value less than the RANGE_MIN_VALUE or more than RANGE_MAX_VALUE. |
|
65
|
|
|
# Decided against raising a ValueError as it is manageable. Should not lead to |
|
66
|
|
|
# unexpected behavior. |
|
67
|
|
|
if boundary_value < self.RANGE_MIN_VALUE: |
|
68
|
|
|
return self.RANGE_MIN_VALUE |
|
69
|
|
|
if boundary_value > self.RANGE_MAX_VALUE: |
|
70
|
|
|
return self.RANGE_MAX_VALUE |
|
71
|
|
|
return boundary_value |
|
72
|
|
|
|
|
73
|
|
|
|
|
74
|
|
|
class HashPartitioner(DefaultPartitioner): |
|
75
|
|
|
|
|
76
|
|
|
def __init__(self, sensor_node_name, hash_ranges): |
|
77
|
|
|
super(HashPartitioner, self).__init__(sensor_node_name=sensor_node_name) |
|
78
|
|
|
self._hash_ranges = self._create_hash_ranges(hash_ranges) |
|
79
|
|
|
|
|
80
|
|
|
def is_sensor_owner(self, sensor_db): |
|
81
|
|
|
return self._is_in_hash_range(sensor_db.get_reference().ref) |
|
82
|
|
|
|
|
83
|
|
|
def get_sensors(self): |
|
84
|
|
|
all_enabled_sensors = get_all_enabled_sensors() |
|
85
|
|
|
|
|
86
|
|
|
partition_members = [] |
|
87
|
|
|
|
|
88
|
|
|
for sensor in all_enabled_sensors: |
|
89
|
|
|
sensor_ref = sensor.get_reference() |
|
90
|
|
|
if self._is_in_hash_range(sensor_ref.ref): |
|
91
|
|
|
partition_members.append(sensor) |
|
92
|
|
|
|
|
93
|
|
|
return partition_members |
|
94
|
|
|
|
|
95
|
|
|
def _is_in_hash_range(self, sensor_ref): |
|
96
|
|
|
sensor_ref_hash = self._hash_sensor_ref(sensor_ref) |
|
97
|
|
|
for hash_range in self._hash_ranges: |
|
98
|
|
|
if sensor_ref_hash in hash_range: |
|
99
|
|
|
return True |
|
100
|
|
|
return False |
|
101
|
|
|
|
|
102
|
|
|
def _hash_sensor_ref(self, sensor_ref): |
|
103
|
|
|
# Hmm... maybe this should be done in C. If it becomes a performance |
|
104
|
|
|
# bottleneck will look at that optimization. |
|
105
|
|
|
|
|
106
|
|
|
# From http://www.cs.hmc.edu/~geoff/classes/hmc.cs070.200101/homework10/hashfuncs.html |
|
107
|
|
|
# The 'liberal' use of ctypes.c_unit is to guarantee unsigned integer and workaround |
|
108
|
|
|
# inifinite precision. |
|
109
|
|
|
md5_hash = hashlib.md5(sensor_ref.encode()) |
|
110
|
|
|
md5_hash_int_repr = int(md5_hash.hexdigest(), 16) |
|
111
|
|
|
h = ctypes.c_uint(0) |
|
112
|
|
|
for d in reversed(str(md5_hash_int_repr)): |
|
113
|
|
|
d = ctypes.c_uint(int(d)) |
|
114
|
|
|
higherorder = ctypes.c_uint(h.value & 0xf8000000) |
|
115
|
|
|
h = ctypes.c_uint(h.value << 5) |
|
116
|
|
|
h = ctypes.c_uint(h.value ^ (higherorder.value >> 27)) |
|
117
|
|
|
h = ctypes.c_uint(h.value ^ d.value) |
|
118
|
|
|
return h.value |
|
119
|
|
|
|
|
120
|
|
|
def _create_hash_ranges(self, hash_ranges_repr): |
|
121
|
|
|
""" |
|
122
|
|
|
Extract from a format like - 0..1024|2048..4096|4096..MAX |
|
123
|
|
|
""" |
|
124
|
|
|
hash_ranges = [] |
|
125
|
|
|
# Likely all this splitting can be avoided and done nicely with regex but I generally |
|
126
|
|
|
# dislike using regex so I go with naive approaches. |
|
127
|
|
|
for range_repr in hash_ranges_repr.split(SUB_RANGE_SEPARATOR): |
|
128
|
|
|
hash_range = Range(range_repr.strip()) |
|
129
|
|
|
hash_ranges.append(hash_range) |
|
130
|
|
|
return hash_ranges |
|
131
|
|
|
|
StackStorm /
st2
