wonderdog 0.1.1 → 0.2.0
Sign up to get free protection for your applications and to get access to all the features.
- data/.gitignore +2 -0
- data/.idea/encodings.xml +5 -0
- data/.idea/misc.xml +5 -0
- data/.idea/modules.xml +9 -0
- data/.idea/scopes/scope_settings.xml +5 -0
- data/.idea/vcs.xml +7 -0
- data/.idea/wonderdog.iml +41 -0
- data/Gemfile +1 -1
- data/bin/estool +22 -1
- data/bin/squirrel.rb +108 -0
- data/lib/wonderdog.rb +3 -0
- data/lib/wonderdog/hadoop_invocation_override.rb +4 -1
- data/lib/wonderdog/version.rb +1 -1
- data/pom.xml +1 -1
- data/spec/spec_helper.rb +1 -1
- data/spec/wonderdog/hadoop_invocation_override_spec.rb +1 -1
- data/squirrel/all_facets.rb +95 -0
- data/squirrel/change_es_index_settings.rb +19 -0
- data/squirrel/clear_es_caches.rb +30 -0
- data/squirrel/esbackup.rb +184 -0
- data/squirrel/esbackup_stripped.rb +153 -0
- data/squirrel/fields.sh +5 -0
- data/squirrel/getFields.rb +19 -0
- data/squirrel/replay.rb +219 -0
- data/squirrel/squirrel.rb +95 -0
- data/squirrel/warmer_interface.rb +59 -0
- data/src/main/java/com/infochimps/elasticsearch/ElasticSearchInputFormat.java +2 -2
- data/src/main/java/com/infochimps/elasticsearch/ElasticSearchStreamingInputFormat.java +14 -2
- data/src/main/java/com/infochimps/elasticsearch/ElasticSearchStreamingOutputFormat.java +20 -5
- data/src/main/java/com/infochimps/elasticsearch/ElasticSearchStreamingRecordReader.java +55 -26
- data/src/main/java/com/infochimps/elasticsearch/ElasticSearchStreamingRecordWriter.java +59 -22
- data/test/cardinality.rb +43 -0
- data/test/change_es_index_settings.rb +19 -0
- data/test/clear_es_caches.rb +30 -0
- data/test/config/mapping.yml +327 -0
- data/test/config/mappings.yml +328 -0
- data/test/count_check.txt +0 -0
- data/test/esbackup_stripped.rb +153 -0
- data/test/mapping.yml +327 -0
- data/test/medium_slow_queries +41 -0
- data/test/queries.txt +0 -0
- data/test/quick_test_slow_queries +4 -0
- data/test/run_pry.rb +3 -0
- data/test/some_slow_queries +53 -0
- data/test/warmer_interface.rb +64 -0
- data/test/warmindices.rb +65 -0
- data/wonderdog.gemspec +1 -1
- metadata +40 -7
@@ -0,0 +1,184 @@
|
|
1
|
+
#!/usr/bin/env ruby
|
2
|
+
|
3
|
+
# Simple script to dump elasticsearch indexes as raw JSON
|
4
|
+
|
5
|
+
require 'tire'
|
6
|
+
require 'zlib'
|
7
|
+
require 'socket'
|
8
|
+
require 'pathname'
|
9
|
+
require 'configliere'
|
10
|
+
require 'multi_json'
|
11
|
+
|
12
|
+
Settings.use :commandline
|
13
|
+
def Settings.available_commands() %w[ backup restore ] ; end
|
14
|
+
def Settings.usage() 'usage: esbackup command [ path ] [..--param=value..]' ; end
|
15
|
+
Settings.description = <<-DESC.gsub(/^ {2}/, '').chomp
|
16
|
+
Simple backup and restore tool for Elasticsearch.
|
17
|
+
|
18
|
+
# Example backup command
|
19
|
+
$ esbackup backup -c localhost -p 9200 -b 100 -i my_index -q '{"query":{"term":{"field":"search_term"}}}'
|
20
|
+
|
21
|
+
# Example restore command
|
22
|
+
$ esbackup restore -c localhost -p 9200 -i my_index -m mapping_file.json
|
23
|
+
|
24
|
+
CAVEAT: Due to stupid fucking restrictions in the gem we use (tire), if you would like to use a dynamic query
|
25
|
+
when backing up you have to specify it like so: -q '{"query_method":["*args_to_method"]}'
|
26
|
+
|
27
|
+
Available commands:
|
28
|
+
#{Settings.available_commands.map{ |cmd| ' ' << cmd }.join("\n")}
|
29
|
+
DESC
|
30
|
+
Settings.define :host, default: Socket.gethostname, flag: 'c', description: 'Host to connect to Elasticsearch'
|
31
|
+
Settings.define :port, default: 9200, flag: 'p', description: 'Port to connect to Elasticsearch'
|
32
|
+
Settings.define :batch_size, default: 10000, flag: 'b', description: 'Batch size'
|
33
|
+
Settings.define :index, flag: 'i', description: 'Index to backup/restore', required: true
|
34
|
+
Settings.define :mappings, flag: 'm', description: 'Dump mappings to or load mappings from provided file'
|
35
|
+
Settings.define :query, flag: 'q', description: 'A JSON hash containing a query to apply to the command (backup only)'
|
36
|
+
Settings.define :dump_file, default: nil flag: 'd', description: 'The name of the JSON dump file, if not passed the index name will be the dump files name'
|
37
|
+
Settings.resolve!
|
38
|
+
|
39
|
+
Tire::Configuration.url "http://#{Settings[:host]}:#{Settings[:port]}"
|
40
|
+
|
41
|
+
class ESBackup
|
42
|
+
|
43
|
+
def initialize(output_dir, options = {})
|
44
|
+
options = options.merge()
|
45
|
+
@output_dir = output_dir || ''
|
46
|
+
@index = options[:index]
|
47
|
+
@batch_size = options[:batch_size].to_i
|
48
|
+
@mapping_file = options[:mappings]
|
49
|
+
@query = MultiJson.load(options[:query]) rescue nil
|
50
|
+
if options[:dump_file].nil?
|
51
|
+
@dump_file = @index
|
52
|
+
else
|
53
|
+
@dump_file = options[:dump_file]
|
54
|
+
end
|
55
|
+
end
|
56
|
+
|
57
|
+
def dump_mapping
|
58
|
+
index = Tire::Index.new @index
|
59
|
+
File.open(@mapping_file, 'w'){ |f| f.puts index.mapping.to_json }
|
60
|
+
end
|
61
|
+
|
62
|
+
def fullpath dir
|
63
|
+
basedir = dir.start_with?('/') ? dir : File.join(Dir.pwd, dir)
|
64
|
+
FileUtils.mkdir_p(basedir)
|
65
|
+
basedir
|
66
|
+
end
|
67
|
+
|
68
|
+
def gz_output
|
69
|
+
File.join(fullpath(@output_dir), @index + '.gz')
|
70
|
+
end
|
71
|
+
|
72
|
+
def create_scanner
|
73
|
+
scan_opts = { size: @batch_size }
|
74
|
+
additional_query = @query
|
75
|
+
Tire::Search::Scan.new(@index, scan_opts) do
|
76
|
+
# This is fucking stupid; why people have to be cute and make everything DSL only
|
77
|
+
# I'll never understand, but the person who wrote this gem has forced us to ONLY be able to
|
78
|
+
# ask queries in this manner.
|
79
|
+
query do
|
80
|
+
additional_query.each_pair do |key, vals|
|
81
|
+
case vals
|
82
|
+
# Assuming here that you are only asking for one field at a time...this is getting hacky fast
|
83
|
+
when Hash then self.send(key.to_sym, *vals.to_a.flatten)
|
84
|
+
when Array then self.send(key.to_sym, *vals)
|
85
|
+
end
|
86
|
+
end
|
87
|
+
end if additional_query
|
88
|
+
end
|
89
|
+
end
|
90
|
+
|
91
|
+
def run
|
92
|
+
dump_mapping if @mapping_file
|
93
|
+
gz = Zlib::GzipWriter.open gz_output
|
94
|
+
count = 0
|
95
|
+
create_scanner.each do |document|
|
96
|
+
document.each do |record|
|
97
|
+
json_doc = record.to_hash.except(:type, :_index, :_explanation, :_score, :_version, :highlight, :sort).to_json
|
98
|
+
gz.puts json_doc
|
99
|
+
count += 1
|
100
|
+
end
|
101
|
+
end
|
102
|
+
gz.close
|
103
|
+
puts "#{@index} backup complete. #{count} records written"
|
104
|
+
end
|
105
|
+
end
|
106
|
+
|
107
|
+
class ESRestore
|
108
|
+
|
109
|
+
def initialize(input, options = {})
|
110
|
+
@index = options[:index]
|
111
|
+
@batch_size = options[:batch_size].to_i
|
112
|
+
@gz_input = Zlib::GzipReader.open(input)
|
113
|
+
@mapping_file = options[:mappings]
|
114
|
+
end
|
115
|
+
|
116
|
+
def create_index
|
117
|
+
index = Tire::Index.new @index
|
118
|
+
options = @mapping_file ? { mappings: MultiJson.load(File.read(@mapping_file)) } : {}
|
119
|
+
index.create(options) unless index.exists?
|
120
|
+
index
|
121
|
+
end
|
122
|
+
|
123
|
+
def run
|
124
|
+
reindex = create_index
|
125
|
+
count, documents = 0, []
|
126
|
+
@gz_input.each_line do |json|
|
127
|
+
documents << MultiJson.load(json)
|
128
|
+
count += 1
|
129
|
+
if count % @batch_size == 0
|
130
|
+
reindex.bulk_create documents
|
131
|
+
puts "#{count} records loaded"
|
132
|
+
documents.clear
|
133
|
+
end
|
134
|
+
end
|
135
|
+
@gz_input.close()
|
136
|
+
reindex.bulk_create documents if not documents.empty?
|
137
|
+
puts "#{@index} restore complete with #{count} records loaded"
|
138
|
+
end
|
139
|
+
end
|
140
|
+
|
141
|
+
class ESDup
|
142
|
+
|
143
|
+
def initialize(input, options = {})
|
144
|
+
@index = options[:index]
|
145
|
+
@batch_size = options[:batch_size].to_i
|
146
|
+
@gz_input = Zlib::GzipReader.open(input)
|
147
|
+
@mapping_file = options[:mappings]
|
148
|
+
end
|
149
|
+
|
150
|
+
def create_index
|
151
|
+
index = Tire::Index.new @index
|
152
|
+
options = @mapping_file ? { mappings: MultiJson.load(File.read(@mapping_file)) } : {}
|
153
|
+
index.create(options) unless index.exists?
|
154
|
+
index
|
155
|
+
end
|
156
|
+
|
157
|
+
def run
|
158
|
+
reindex = create_index
|
159
|
+
count, documents = 0, []
|
160
|
+
@gz_input.each_line do |json|
|
161
|
+
line = MultiJson.load(json)
|
162
|
+
line.delete("_id")
|
163
|
+
line.delete("id")
|
164
|
+
documents << line
|
165
|
+
count += 1
|
166
|
+
if count % @batch_size == 0
|
167
|
+
reindex.bulk_create documents
|
168
|
+
puts "#{count} records loaded"
|
169
|
+
documents.clear
|
170
|
+
end
|
171
|
+
end
|
172
|
+
@gz_input.close()
|
173
|
+
reindex.bulk_create documents if not documents.empty?
|
174
|
+
puts "#{@index} restore complete with #{count} records loaded"
|
175
|
+
end
|
176
|
+
end
|
177
|
+
|
178
|
+
case command = Settings.rest.shift.to_s.to_sym
|
179
|
+
when :restore then ESRestore.new(Settings.rest.shift, Settings.to_hash).run
|
180
|
+
when :backup then ESBackup.new(Settings.rest.shift, Settings.to_hash).run
|
181
|
+
when :duplicate then ESDup.new(Settings.rest.shift, Settings.to_hash).run
|
182
|
+
else abort Settings.help("Must specify either backup, restore or duplicate. Got <#{command}>")
|
183
|
+
end
|
184
|
+
|
@@ -0,0 +1,153 @@
|
|
1
|
+
#!/usr/bin/env ruby
|
2
|
+
|
3
|
+
# Simple script to dump elasticsearch indexes as raw JSON
|
4
|
+
|
5
|
+
require 'tire'
|
6
|
+
require 'zlib'
|
7
|
+
require 'socket'
|
8
|
+
require 'pathname'
|
9
|
+
require 'multi_json'
|
10
|
+
|
11
|
+
class ESBackup
|
12
|
+
|
13
|
+
def initialize(output_dir, options = {})
|
14
|
+
Tire::Configuration.url "http://#{options[:host]}:#{options[:port]}"
|
15
|
+
@output_dir = output_dir || ''
|
16
|
+
@index = options[:index]
|
17
|
+
@batch_size = options[:batch_size].to_i
|
18
|
+
@mapping_file = options[:mappings]
|
19
|
+
if options[:query].nil?
|
20
|
+
@query = nil
|
21
|
+
else
|
22
|
+
@query = MultiJson.load(options[:query]) rescue nil
|
23
|
+
end
|
24
|
+
if options[:dump_file].nil?
|
25
|
+
@dump_file = @index
|
26
|
+
else
|
27
|
+
@dump_file = options[:dump_file]
|
28
|
+
end
|
29
|
+
end
|
30
|
+
|
31
|
+
def dump_mapping
|
32
|
+
index = Tire::Index.new @index
|
33
|
+
File.open(@mapping_file, 'w'){ |f| f.puts index.mapping.to_json }
|
34
|
+
end
|
35
|
+
|
36
|
+
def fullpath dir
|
37
|
+
basedir = dir.start_with?('/') ? dir : File.join(Dir.pwd, dir)
|
38
|
+
FileUtils.mkdir_p(basedir)
|
39
|
+
basedir
|
40
|
+
end
|
41
|
+
|
42
|
+
def gz_output
|
43
|
+
File.join(fullpath(@output_dir), @index + '.gz')
|
44
|
+
end
|
45
|
+
|
46
|
+
def create_scanner
|
47
|
+
scan_opts = { size: @batch_size }
|
48
|
+
additional_query = @query
|
49
|
+
Tire::Search::Scan.new(@index, scan_opts) do
|
50
|
+
# This is fucking stupid; why people have to be cute and make everything DSL only
|
51
|
+
# I'll never understand, but the person who wrote this gem has forced us to ONLY be able to
|
52
|
+
# ask queries in this manner.
|
53
|
+
query do
|
54
|
+
additional_query.each_pair do |key, vals|
|
55
|
+
case vals
|
56
|
+
# Assuming here that you are only asking for one field at a time...this is getting hacky fast
|
57
|
+
when Hash then self.send(key.to_sym, *vals.to_a.flatten)
|
58
|
+
when Array then self.send(key.to_sym, *vals)
|
59
|
+
end
|
60
|
+
end
|
61
|
+
end if additional_query
|
62
|
+
end
|
63
|
+
end
|
64
|
+
|
65
|
+
def run
|
66
|
+
dump_mapping if @mapping_file
|
67
|
+
gz = Zlib::GzipWriter.open gz_output
|
68
|
+
count = 0
|
69
|
+
create_scanner.each do |document|
|
70
|
+
document.each do |record|
|
71
|
+
json_doc = record.to_hash.except(:type, :_index, :_explanation, :_score, :_version, :highlight, :sort).to_json
|
72
|
+
gz.puts json_doc
|
73
|
+
count += 1
|
74
|
+
end
|
75
|
+
end
|
76
|
+
gz.close
|
77
|
+
puts "#{@index} backup complete. #{count} records written"
|
78
|
+
end
|
79
|
+
end
|
80
|
+
|
81
|
+
class ESRestore
|
82
|
+
|
83
|
+
def initialize(input, options = {})
|
84
|
+
Tire::Configuration.url "http://#{options[:host]}:#{options[:port]}"
|
85
|
+
@index = options[:index]
|
86
|
+
@batch_size = options[:batch_size].to_i
|
87
|
+
@gz_input = Zlib::GzipReader.open(input)
|
88
|
+
@mapping_file = options[:mappings]
|
89
|
+
end
|
90
|
+
|
91
|
+
def create_index
|
92
|
+
index = Tire::Index.new @index
|
93
|
+
options = @mapping_file ? { mappings: MultiJson.load(File.read(@mapping_file)) } : {}
|
94
|
+
index.create(options) unless index.exists?
|
95
|
+
index
|
96
|
+
end
|
97
|
+
|
98
|
+
def run
|
99
|
+
reindex = create_index
|
100
|
+
count, documents = 0, []
|
101
|
+
@gz_input.each_line do |json|
|
102
|
+
documents << MultiJson.load(json)
|
103
|
+
count += 1
|
104
|
+
if count % @batch_size == 0
|
105
|
+
reindex.bulk_create documents
|
106
|
+
puts "#{count} records loaded"
|
107
|
+
documents.clear
|
108
|
+
end
|
109
|
+
end
|
110
|
+
@gz_input.close()
|
111
|
+
reindex.bulk_create documents if not documents.empty?
|
112
|
+
puts "#{@index} restore complete with #{count} records loaded"
|
113
|
+
end
|
114
|
+
end
|
115
|
+
|
116
|
+
class ESDup
|
117
|
+
|
118
|
+
def initialize(input, options = {})
|
119
|
+
Tire::Configuration.url "http://#{options[:host]}:#{options[:port]}"
|
120
|
+
@index = options[:index]
|
121
|
+
@batch_size = options[:batch_size].to_i
|
122
|
+
@gz_input = Zlib::GzipReader.open(input)
|
123
|
+
@mapping_file = options[:mappings]
|
124
|
+
end
|
125
|
+
|
126
|
+
def create_index
|
127
|
+
index = Tire::Index.new @index
|
128
|
+
options = @mapping_file ? { mappings: MultiJson.load(File.read(@mapping_file)) } : {}
|
129
|
+
index.create(options) unless index.exists?
|
130
|
+
index
|
131
|
+
end
|
132
|
+
|
133
|
+
def run
|
134
|
+
reindex = create_index
|
135
|
+
count, documents = 0, []
|
136
|
+
@gz_input.each_line do |json|
|
137
|
+
line = MultiJson.load(json)
|
138
|
+
line.delete("_id")
|
139
|
+
line.delete("id")
|
140
|
+
documents << line
|
141
|
+
count += 1
|
142
|
+
if count % @batch_size == 0
|
143
|
+
reindex.bulk_create documents
|
144
|
+
puts "#{count} records loaded"
|
145
|
+
documents.clear
|
146
|
+
end
|
147
|
+
end
|
148
|
+
@gz_input.close()
|
149
|
+
reindex.bulk_create documents if not documents.empty?
|
150
|
+
puts "#{@index} restore complete with #{count} records loaded"
|
151
|
+
end
|
152
|
+
end
|
153
|
+
|
data/squirrel/fields.sh
ADDED
@@ -0,0 +1,5 @@
|
|
1
|
+
for foo in flight_id metric tb_h feature base_feature metric_feature cnt; do
|
2
|
+
echo $foo;
|
3
|
+
/home/maphysics/.rbenv/shims/ruby /home/maphysics/GitProjects/wonderdog/test/getFields.rb --dump=/home/maphysics/GitProjects/wonderdog/test/flight_count_20130405 --field=$foo >> $foo.txt ;
|
4
|
+
cat $foo.txt |sort | uniq -c |sort -n | wc -l;
|
5
|
+
done
|
@@ -0,0 +1,19 @@
|
|
1
|
+
require 'rubygems'
|
2
|
+
require 'configliere'
|
3
|
+
require 'json'
|
4
|
+
require 'multi_json'
|
5
|
+
|
6
|
+
Settings.use :commandline
|
7
|
+
Settings.use :config_block
|
8
|
+
Settings.define :dump
|
9
|
+
Settings.define :field
|
10
|
+
Settings.resolve!
|
11
|
+
|
12
|
+
def get_value_counts(dump, field)
|
13
|
+
File.open(dump).each do |line|
|
14
|
+
record = MultiJson.load(line)
|
15
|
+
puts record[field]
|
16
|
+
end
|
17
|
+
end
|
18
|
+
|
19
|
+
get_value_counts(Settings.dump, Settings.field)
|
data/squirrel/replay.rb
ADDED
@@ -0,0 +1,219 @@
|
|
1
|
+
#!/usr/bin/env ruby
|
2
|
+
|
3
|
+
require 'rubygems'
|
4
|
+
require 'json'
|
5
|
+
require 'time'
|
6
|
+
|
7
|
+
|
8
|
+
#require 'random'
|
9
|
+
|
10
|
+
########################################################################################################################
|
11
|
+
# This program is designed to read an elasticsearch log file and return #
|
12
|
+
# information about how long a slow process took, run the query, and #
|
13
|
+
# return information about how long it took to run the query again. #
|
14
|
+
# Example command: #
|
15
|
+
# ruby ./replay.rb --logfile=/var/log/elasticsearch/patrick.log --port=9200 --host=localhost #
|
16
|
+
########################################################################################################################
|
17
|
+
|
18
|
+
########################################################################################################################
|
19
|
+
# Global variables for storing metadata #
|
20
|
+
########################################################################################################################
|
21
|
+
@slowlog_lines = []
|
22
|
+
@metadata_hash = {}
|
23
|
+
|
24
|
+
########################################################################################################################
|
25
|
+
# Parse logfile, grab: #
|
26
|
+
# *the timestamp #
|
27
|
+
# *the index #
|
28
|
+
# *the node #
|
29
|
+
# *the type of search #
|
30
|
+
# *the time in millisecond #
|
31
|
+
# *At least first 50 char of query #
|
32
|
+
########################################################################################################################
|
33
|
+
|
34
|
+
class ParseMetaData
|
35
|
+
attr_accessor :metaData
|
36
|
+
|
37
|
+
def initialize(metaString, metaArray = [])
|
38
|
+
@metaString = metaString
|
39
|
+
@metaArray = metaArray
|
40
|
+
@metaData = {}
|
41
|
+
@bracket_pairs = get_bracket_pair_indexes
|
42
|
+
end
|
43
|
+
|
44
|
+
def get_left_bracket_indexes
|
45
|
+
@metaString.enum_for(:scan, Regexp.new('\[')).map {Regexp.last_match.begin(0)}
|
46
|
+
end
|
47
|
+
|
48
|
+
def get_right_bracket_indexes
|
49
|
+
@metaString.enum_for(:scan, Regexp.new('\]')).map {Regexp.last_match.begin(0)}
|
50
|
+
end
|
51
|
+
|
52
|
+
def get_bracket_pair_indexes
|
53
|
+
get_left_bracket_indexes.zip(get_right_bracket_indexes)
|
54
|
+
end
|
55
|
+
|
56
|
+
def get_query
|
57
|
+
startInd = @metaString.enum_for(:scan, Regexp.new(' source\[')).map {Regexp.last_match.begin(0)+8}
|
58
|
+
endInd = @metaString.enum_for(:scan, Regexp.new('_source\[')).map {Regexp.last_match.begin(0)-9}
|
59
|
+
@metaData["query"] = @metaString[startInd[0]..endInd[0]]
|
60
|
+
end
|
61
|
+
|
62
|
+
def find_meta_data(meta)
|
63
|
+
start = @metaString.enum_for(:scan, Regexp.new(meta)).map {Regexp.last_match.begin(0) + meta.size}
|
64
|
+
index = get_left_bracket_indexes.index(start[0])
|
65
|
+
unless index.nil?
|
66
|
+
bracket_pair = @bracket_pairs[index]
|
67
|
+
#puts @metaString[bracket_pair[0]+1..bracket_pair[1]-1].inspect
|
68
|
+
@metaData[meta] = @metaString[bracket_pair[0]+1..bracket_pair[1]-1]
|
69
|
+
end
|
70
|
+
end
|
71
|
+
|
72
|
+
def get_extra_meta_data
|
73
|
+
@metaArray.each do |meta|
|
74
|
+
find_meta_data(meta)
|
75
|
+
end
|
76
|
+
end
|
77
|
+
|
78
|
+
def get_basic_meta_data
|
79
|
+
#FIXME! Make this dynamic and depended on the first four [] to contain the same things everytime
|
80
|
+
@metaData["timestamp"] = @metaString[@bracket_pairs[0][0]+1..@bracket_pairs[0][1]-1]
|
81
|
+
@metaData["node"] = @metaString[@bracket_pairs[3][0]+1..@bracket_pairs[3][1]-1]
|
82
|
+
@metaData["index"] = @metaString[@bracket_pairs[4][0]+1..@bracket_pairs[4][1]-1]
|
83
|
+
end
|
84
|
+
|
85
|
+
def get_meta_data
|
86
|
+
get_basic_meta_data
|
87
|
+
get_query
|
88
|
+
unless @metaArray == []
|
89
|
+
get_extra_meta_data
|
90
|
+
end
|
91
|
+
end
|
92
|
+
end
|
93
|
+
|
94
|
+
def parse_logline(line, metaArray)
|
95
|
+
|
96
|
+
if (line =~ %r{, source\[(.*)\], extra_source})
|
97
|
+
query = $1
|
98
|
+
else
|
99
|
+
warn("couldn't parse line")
|
100
|
+
return
|
101
|
+
end
|
102
|
+
|
103
|
+
#puts line
|
104
|
+
parser = ParseMetaData.new(line, metaArray)
|
105
|
+
parser.get_meta_data
|
106
|
+
|
107
|
+
return parser.metaData["query"], parser.metaData
|
108
|
+
end
|
109
|
+
|
110
|
+
########################################################################################################################
|
111
|
+
# Return the following info to stdout as tab delimited: #
|
112
|
+
# Current time #
|
113
|
+
# Original timestamp #
|
114
|
+
# Duration of query in log #
|
115
|
+
# Duration of re-ran query according to elastic search #
|
116
|
+
# Duration of re-ran query according to the wall clock #
|
117
|
+
# The meta captured from the logfile #
|
118
|
+
# A snippet of query #
|
119
|
+
# Extra source data from logfile #
|
120
|
+
########################################################################################################################
|
121
|
+
class Replay
|
122
|
+
|
123
|
+
def initialize(logfile, host, port, preference, routing)
|
124
|
+
@logfile = logfile
|
125
|
+
@host = host
|
126
|
+
@port = port
|
127
|
+
@preference = preference
|
128
|
+
@routing = routing
|
129
|
+
end
|
130
|
+
|
131
|
+
def header()
|
132
|
+
puts "\n"
|
133
|
+
puts %w[current_timestamp original_timestamp es_duration(ms) new_duration(ms) clock_time_duration(ms) node index query_fragment].join("\t")
|
134
|
+
end
|
135
|
+
|
136
|
+
def output(query, data, malformed=false)
|
137
|
+
query_fragment = query[0..49]
|
138
|
+
if malformed
|
139
|
+
puts "malformed"
|
140
|
+
puts query_fragment
|
141
|
+
else
|
142
|
+
took = data['took'].to_s
|
143
|
+
current_time = data['new_timestamp'].to_s
|
144
|
+
original_timestamp = data['timestamp'].to_s
|
145
|
+
es_duration = data['original_dur'].to_s
|
146
|
+
new_duration = data['new_duration'].to_i.to_s
|
147
|
+
node = data['node'].to_s
|
148
|
+
index = data['index'].to_s
|
149
|
+
if Random.rand() < 0.1
|
150
|
+
header
|
151
|
+
end
|
152
|
+
puts [current_time, original_timestamp, es_duration, took, new_duration, node, index, query_fragment].join("\t")
|
153
|
+
end
|
154
|
+
end
|
155
|
+
|
156
|
+
def build_curl_command_string(query, data)
|
157
|
+
base_uri = "'#{@host}:#{@port}/#{data['index']}/_search"
|
158
|
+
if @preference[0] && @routing[0]
|
159
|
+
base_uri.concat("?preference=#{@preference[1]},routing=#{@routing[1]}")
|
160
|
+
elsif @preference[0] && !@routing[0]
|
161
|
+
base_uri.concat("?reference=#{@preference[1]}")
|
162
|
+
elsif @routing[0] && !@preference[0]
|
163
|
+
base_uri.concat("routing=#{@routing[1]}")
|
164
|
+
end
|
165
|
+
curl_command = "curl -s -XGET ".concat(base_uri)
|
166
|
+
curl_command.concat("/' -d '#{query}'")
|
167
|
+
end
|
168
|
+
|
169
|
+
########################################################################################################################
|
170
|
+
# Execute slow query from log #
|
171
|
+
########################################################################################################################
|
172
|
+
|
173
|
+
def execute_query(total_took, query, data)
|
174
|
+
if query.include? " " or query.index('(\\\'.*?\\\')').nil?
|
175
|
+
if data['search_type'] == "QUERY_THEN_FETCH"
|
176
|
+
data['new_timestamp'] = Time.now
|
177
|
+
data['new_start_time'] = Time.now.to_f * 1000
|
178
|
+
cmd = build_curl_command_string(query, data)
|
179
|
+
#puts cmd
|
180
|
+
curl_result = `#{cmd}`
|
181
|
+
#puts curl_result
|
182
|
+
#puts "\n"
|
183
|
+
data['new_end_time'] = Time.now.to_f * 1000
|
184
|
+
data['new_duration'] = data['new_end_time'] - data['new_start_time']
|
185
|
+
data['original_dur'] = data['took']
|
186
|
+
data = data.merge(JSON.parse(curl_result))
|
187
|
+
output(query, data)
|
188
|
+
else
|
189
|
+
puts "error don't know search type, please throw an exception here"
|
190
|
+
end
|
191
|
+
else
|
192
|
+
puts "malformed query string"
|
193
|
+
puts query
|
194
|
+
output(query, data, malformed=true)
|
195
|
+
end
|
196
|
+
total_took + data['new_duration'].to_i
|
197
|
+
end
|
198
|
+
|
199
|
+
########################################################################################################################
|
200
|
+
# MAIN #
|
201
|
+
########################################################################################################################
|
202
|
+
|
203
|
+
def run
|
204
|
+
sl_regex = Regexp.new(('(slowlog\\.query)'), Regexp::IGNORECASE)
|
205
|
+
metaArray = %w[took took_millis types search_type total_shards]
|
206
|
+
header
|
207
|
+
total_took = 0
|
208
|
+
File.readlines(@logfile).each do |line|
|
209
|
+
if sl_regex.match(line)
|
210
|
+
query, query_hash = parse_logline(line, metaArray)
|
211
|
+
total_took = execute_query(total_took, query, query_hash)
|
212
|
+
end
|
213
|
+
end
|
214
|
+
total_took /= 60000.0
|
215
|
+
puts "All together the slow logs took: #{total_took}min"
|
216
|
+
end
|
217
|
+
end
|
218
|
+
|
219
|
+
|