Edited at

Chef-Zeroのキャパシティが気になったのでRubyでベンチマークをとった

More than 3 years have passed since last update.


この記事は最終更新から1年以上経過しています。 気をつけてね。


実行したコードはこちら。 higanworks/chefzero-benchmark

The Ruby Toolbox - Benchmarkingからなにかしら結果が見やすいとか使いやすいのはないかと見てみたけど、結局普通のBenchmark.measureを使い、次のやり方で計測しました。


  1. Chef-Zefoを起動する

  2. ダミーのノードをFauxhaiで生成する


  3. RidleyでノードをChef-Zeroに登録する


    • 登録数は [50, 100, 500, 1000, 5000] の5パターン



  4. 次の3つを計測


    • ノードを1つ抽出する

    • サーチで単純なクエリにマッチするNodeを抽出する

    • サーチでワイルドカードを含むクエリにマッチするNodeを抽出する



Chef-Zeroの保存方式(データストア)は標準のオンメモリと、実際のディレクトリを直接使用するファイルシステム(※こっちに期待)の2種類。

ベンチマークはこんな条件の下で行った。

MacBook Pro Retina, 13-inch, Late 2013

Processor: 2.8 GHz Intel Core i7
Memory: 16 GB 1600 MHz DDR3

RUBY_VERSION #=> 2.1.2
RbConfig::CONFIG["PATCHLEVEL"] #=> 95
ChefZero::VERSION #=> 3.0


結果のまとめ

どちらのデータストアでも、単体を取得する分には特に劣化がなかった。

オンメモリなら1000ノードのサーチも2秒程度、流石に5000ノードあったら10秒以上かかっていたけど許容範囲。

本命のファイルシステムだと、500ノードを超えたあたりからサーチが徐々に厳しくなっていく。

5000ノードでサーチすると流石にRidley標準のタイムアウトで終了した。

でもサーチを使わなければ、動かすマシンにもよるが5000程度のオブジェクトがあってもへっちゃら感です。


ベンチマークのソース

Benchmark#measureを使う部分はまとめるべきだったと思いつつベタ書き。


Rakefile

require 'chef_zero/server'

require 'ridley'
require 'fauxhai'
require 'benchmark'
require 'tempfile'
require 'faker'

TIMES = [50, 100, 500, 1000, 5000]

def ridley_init
f = Tempfile.new('pem')
File.write(f.path, OpenSSL::PKey::RSA.new(2048))

ridley = Ridley.new(
server_url: "http://localhost:4000",
client_name: "reset",
client_key: f.path
)
ridley.logger.level = 3
ridley
end

def create_fake_node(ridley, count)
count.times do |t|
obj = ridley.node.new
obj.name = [Faker::Internet.domain_word + t.to_s , Faker::Internet.domain_name].join('.')
@nodes << obj.name
obj.chef_environment = get_env_sample
obj.automatic = Fauxhai.mock(platform: 'ubuntu', version: '12.04').data
ridley.node.create(obj)
end
end

def get_env_sample
[
'development',
'production',
'office',
'resort'
].sample
end

def version
puts '============================'
puts ['RUBY_VERSION', RUBY_VERSION].join(' #=> ')
puts ['RbConfig::CONFIG["PATCHLEVEL"]', RbConfig::CONFIG['PATCHLEVEL']].join(' #=> ')
puts ['ChefZero::VERSION', ChefZero::VERSION].join(' #=> ')
puts '============================'
end

desc 'run benchmark both mem_store and disk_store.'
task :default do
Rake::Task['bench:mem'].invoke
Rake::Task['bench:disk'].invoke
end

desc 'memory_store benchmack'
namespace :bench do
task :mem do
version
server = ChefZero::Server.new(port: 4000)
server.start_background

ridley = ridley_init

TIMES.each do |time|
puts "Preparing fake node objects upto #{time}..."
## Warmup
@nodes = []
puts Benchmark::CAPTION
puts Benchmark.measure {
create_fake_node(ridley, time)
}
puts "Done!"
puts "============================"

puts "Start benchmark for pick up one node 5 times of #{time} nodes (on memory)"
5.times do
puts Benchmark.measure {
arr = ridley.node.find(@nodes.sample)
puts Benchmark::CAPTION
}
end

puts "Start benchmark for search 5 times by simple match query of #{time} nodes (on memory)"
5.times do
puts Benchmark.measure {
arr = ridley.search(:node, "chef_environment:#{get_env_sample}")
puts "--------Matched #{arr.size} nodes of #{time} objects"
puts Benchmark::CAPTION
}
end
puts "Start benchmark for search 5 times with wildcard match query of #{time} nodes (on memory)"
5.times do
puts Benchmark.measure {
arr = ridley.search(:node, "name:*#{Faker::Internet.domain_suffix}")
puts "--------Matched #{arr.size} nodes of #{time} objects"
puts Benchmark::CAPTION
}
end
puts "Done! Data will be cleared."
puts "============================\n\n"
server.clear_data
end
server.stop
end

desc 'disk_store benchmack'
task :disk do
require 'chef/chef_fs/chef_fs_data_store'
require 'chef/chef_fs/config'
require 'chef_zero/data_store/raw_file_store'

version

Dir.mktmpdir do |tmpdir|
puts "Temporary data_store was created to #{tmpdir}."
Chef::Config[:chef_repo_path] = tmpdir
chef_fs = Chef::ChefFS::Config.new.local_fs
chef_fs.write_pretty_json = true
data_store = Chef::ChefFS::ChefFSDataStore.new(chef_fs)

server_options = {}
server_options[:data_store] = data_store
server_options[:port] = 4000

server = ChefZero::Server.new(server_options)
server.start_background

ridley = ridley_init

TIMES.each do |time|
puts "Preparing fake node objects upto #{time}..."
## Warmup
@nodes = []
puts Benchmark::CAPTION
puts Benchmark.measure {
create_fake_node(ridley, time)
}
puts "Done!"
puts "============================"

puts "Start benchmark for pick up one node 5 times of #{time} nodes (on filesystem)"
5.times do
puts Benchmark.measure {
arr = ridley.node.find(@nodes.sample)
puts Benchmark::CAPTION
}
end

puts "Start benchmark for search 5 times by simple match query of #{time} nodes (on filesystem)"
5.times do
puts Benchmark.measure {
arr = ridley.search(:node, "chef_environment:#{get_env_sample}")
puts "--------Matched #{arr.size} nodes of #{time} objects"
puts Benchmark::CAPTION
}
end
puts "Start benchmark for search 5 times with wildcard match query of #{time} nodes (on filesystem)"
5.times do
puts Benchmark.measure {
arr = ridley.search(:node, "name:*#{Faker::Internet.domain_suffix}")
puts "--------Matched #{arr.size} nodes of #{time} objects"
puts Benchmark::CAPTION
}
end

puts "Done! Data will be cleared."
puts "============================\n\n"

chef_fs.child_paths.each_value do |dirs|
dirs.each do |dir|
if File.exists?(dir)
FileUtils.rm(Dir.glob("#{dir}/*.json"))
end
end
end
server.clear_data
end
server.stop
end
end
end



ベンチマークの実行結果

$ rake 

============================
RUBY_VERSION #=> 2.1.2
RbConfig::CONFIG["PATCHLEVEL"] #=> 95
ChefZero::VERSION #=> 3.0
============================
Preparing fake node objects upto 50...
user system total real
0.750000 0.070000 0.820000 ( 1.190110)
Done!
============================
Start benchmark for pick up one node 5 times of 50 nodes (on memory)
user system total real
0.010000 0.000000 0.010000 ( 0.009239)
user system total real
0.010000 0.000000 0.010000 ( 0.009327)
user system total real
0.010000 0.000000 0.010000 ( 0.009776)
user system total real
0.010000 0.000000 0.010000 ( 0.009797)
user system total real
0.000000 0.000000 0.000000 ( 0.009014)
Start benchmark for search 5 times by simple match query of 50 nodes (on memory)
--------Matched 13 nodes of 50 objects
user system total real
0.150000 0.010000 0.160000 ( 0.153010)
--------Matched 16 nodes of 50 objects
user system total real
0.130000 0.010000 0.140000 ( 0.133699)
--------Matched 11 nodes of 50 objects
user system total real
0.120000 0.010000 0.130000 ( 0.132803)
--------Matched 11 nodes of 50 objects
user system total real
0.110000 0.000000 0.110000 ( 0.114197)
--------Matched 10 nodes of 50 objects
user system total real
0.150000 0.010000 0.160000 ( 0.156211)
Start benchmark for search 5 times with wildcard match query of 50 nodes (on memory)
--------Matched 10 nodes of 50 objects
user system total real
0.110000 0.000000 0.110000 ( 0.112973)
--------Matched 9 nodes of 50 objects
user system total real
0.110000 0.010000 0.120000 ( 0.117174)
--------Matched 10 nodes of 50 objects
user system total real
0.110000 0.000000 0.110000 ( 0.120937)
--------Matched 6 nodes of 50 objects
user system total real
0.120000 0.000000 0.120000 ( 0.117076)
--------Matched 6 nodes of 50 objects
user system total real
0.100000 0.010000 0.110000 ( 0.110966)
Done! Data will be cleared.
============================

Preparing fake node objects upto 100...
user system total real
1.170000 0.130000 1.300000 ( 1.319404)
Done!
============================
Start benchmark for pick up one node 5 times of 100 nodes (on memory)
user system total real
0.000000 0.010000 0.010000 ( 0.007963)
user system total real
0.010000 0.000000 0.010000 ( 0.008128)
user system total real
0.010000 0.000000 0.010000 ( 0.008020)
user system total real
0.000000 0.000000 0.000000 ( 0.007828)
user system total real
0.010000 0.000000 0.010000 ( 0.008201)
Start benchmark for search 5 times by simple match query of 100 nodes (on memory)
--------Matched 16 nodes of 100 objects
user system total real
0.250000 0.010000 0.260000 ( 0.258547)
--------Matched 22 nodes of 100 objects
user system total real
0.210000 0.000000 0.210000 ( 0.215202)
--------Matched 22 nodes of 100 objects
user system total real
0.250000 0.010000 0.260000 ( 0.264075)
--------Matched 22 nodes of 100 objects
user system total real
0.230000 0.000000 0.230000 ( 0.239582)
--------Matched 34 nodes of 100 objects
user system total real
0.260000 0.010000 0.270000 ( 0.268535)
Start benchmark for search 5 times with wildcard match query of 100 nodes (on memory)
--------Matched 12 nodes of 100 objects
user system total real
0.230000 0.020000 0.250000 ( 0.240597)
--------Matched 22 nodes of 100 objects
user system total real
0.320000 0.020000 0.340000 ( 0.341606)
--------Matched 14 nodes of 100 objects
user system total real
0.240000 0.000000 0.240000 ( 0.243652)
--------Matched 12 nodes of 100 objects
user system total real
0.190000 0.010000 0.200000 ( 0.197043)
--------Matched 19 nodes of 100 objects
user system total real
0.210000 0.000000 0.210000 ( 0.215121)
Done! Data will be cleared.
============================

Preparing fake node objects upto 500...
user system total real
5.730000 0.660000 6.390000 ( 6.489995)
Done!
============================
Start benchmark for pick up one node 5 times of 500 nodes (on memory)
user system total real
0.010000 0.000000 0.010000 ( 0.008062)
user system total real
0.000000 0.000000 0.000000 ( 0.007949)
user system total real
0.010000 0.000000 0.010000 ( 0.007695)
user system total real
0.010000 0.000000 0.010000 ( 0.007941)
user system total real
0.010000 0.010000 0.020000 ( 0.008567)
Start benchmark for search 5 times by simple match query of 500 nodes (on memory)
--------Matched 141 nodes of 500 objects
user system total real
1.190000 0.020000 1.210000 ( 1.224797)
--------Matched 134 nodes of 500 objects
user system total real
1.260000 0.060000 1.320000 ( 1.316611)
--------Matched 141 nodes of 500 objects
user system total real
1.300000 0.030000 1.330000 ( 1.329287)
--------Matched 134 nodes of 500 objects
user system total real
1.300000 0.030000 1.330000 ( 1.333752)
--------Matched 112 nodes of 500 objects
user system total real
1.330000 0.040000 1.370000 ( 1.369286)
Start benchmark for search 5 times with wildcard match query of 500 nodes (on memory)
--------Matched 88 nodes of 500 objects
user system total real
1.040000 0.020000 1.060000 ( 1.059881)
--------Matched 86 nodes of 500 objects
user system total real
1.200000 0.050000 1.250000 ( 1.254659)
--------Matched 88 nodes of 500 objects
user system total real
1.170000 0.020000 1.190000 ( 1.189626)
--------Matched 86 nodes of 500 objects
user system total real
1.080000 0.020000 1.100000 ( 1.099927)
--------Matched 76 nodes of 500 objects
user system total real
1.240000 0.030000 1.270000 ( 1.264090)
Done! Data will be cleared.
============================

Preparing fake node objects upto 1000...
user system total real
11.460000 1.310000 12.770000 ( 12.823895)
Done!
============================
Start benchmark for pick up one node 5 times of 1000 nodes (on memory)
user system total real
0.010000 0.000000 0.010000 ( 0.007845)
user system total real
0.000000 0.000000 0.000000 ( 0.008117)
user system total real
0.010000 0.000000 0.010000 ( 0.007736)
user system total real
0.010000 0.000000 0.010000 ( 0.007951)
user system total real
0.010000 0.000000 0.010000 ( 0.009608)
Start benchmark for search 5 times by simple match query of 1000 nodes (on memory)
--------Matched 238 nodes of 1000 objects
user system total real
2.460000 0.040000 2.500000 ( 2.514587)
--------Matched 247 nodes of 1000 objects
user system total real
2.540000 0.050000 2.590000 ( 2.595222)
--------Matched 261 nodes of 1000 objects
user system total real
2.750000 0.060000 2.810000 ( 2.810951)
--------Matched 247 nodes of 1000 objects
user system total real
2.550000 0.050000 2.600000 ( 2.621313)
--------Matched 238 nodes of 1000 objects
user system total real
2.650000 0.050000 2.700000 ( 2.719356)
Start benchmark for search 5 times with wildcard match query of 1000 nodes (on memory)
--------Matched 174 nodes of 1000 objects
user system total real
2.540000 0.050000 2.590000 ( 2.600654)
--------Matched 164 nodes of 1000 objects
user system total real
2.850000 0.080000 2.930000 ( 2.945433)
--------Matched 174 nodes of 1000 objects
user system total real
3.050000 0.060000 3.110000 ( 3.134419)
--------Matched 185 nodes of 1000 objects
user system total real
2.520000 0.050000 2.570000 ( 2.582858)
--------Matched 174 nodes of 1000 objects
user system total real
2.410000 0.040000 2.450000 ( 2.453993)
Done! Data will be cleared.
============================

Preparing fake node objects upto 5000...
user system total real
58.740000 6.480000 65.220000 ( 65.438612)
Done!
============================
Start benchmark for pick up one node 5 times of 5000 nodes (on memory)
user system total real
0.010000 0.000000 0.010000 ( 0.008116)
user system total real
0.000000 0.000000 0.000000 ( 0.008034)
user system total real
0.010000 0.000000 0.010000 ( 0.008345)
user system total real
0.010000 0.000000 0.010000 ( 0.008565)
user system total real
0.010000 0.000000 0.010000 ( 0.008128)
Start benchmark for search 5 times by simple match query of 5000 nodes (on memory)
--------Matched 1234 nodes of 5000 objects
user system total real
21.720000 0.440000 22.160000 ( 22.189359)
--------Matched 1234 nodes of 5000 objects
user system total real
16.100000 0.330000 16.430000 ( 16.437395)
--------Matched 1234 nodes of 5000 objects
user system total real
12.600000 0.350000 12.950000 ( 12.960188)
--------Matched 1280 nodes of 5000 objects
user system total real
13.010000 0.320000 13.330000 ( 13.400405)
--------Matched 1280 nodes of 5000 objects
user system total real
13.110000 0.300000 13.410000 ( 13.486141)
Start benchmark for search 5 times with wildcard match query of 5000 nodes (on memory)
--------Matched 852 nodes of 5000 objects
user system total real
13.220000 0.460000 13.680000 ( 13.835438)
--------Matched 862 nodes of 5000 objects
user system total real
12.290000 0.300000 12.590000 ( 12.635503)
--------Matched 871 nodes of 5000 objects
user system total real
13.030000 0.250000 13.280000 ( 13.518441)
--------Matched 862 nodes of 5000 objects
user system total real
12.390000 0.240000 12.630000 ( 12.693962)
--------Matched 813 nodes of 5000 objects
user system total real
11.980000 0.220000 12.200000 ( 12.211569)
Done! Data will be cleared.
============================

============================
RUBY_VERSION #=> 2.1.2
RbConfig::CONFIG["PATCHLEVEL"] #=> 95
ChefZero::VERSION #=> 3.0
============================
Temporary data_store was created to /var/folders/s9/kwcs_8ln0d32v_n4hdfb7td00000gn/T/d20140826-31513-1innb9o.
Preparing fake node objects upto 50...
user system total real
0.690000 0.090000 0.780000 ( 0.795218)
Done!
============================
Start benchmark for pick up one node 5 times of 50 nodes (on filesystem)
user system total real
0.010000 0.000000 0.010000 ( 0.007690)
user system total real
0.000000 0.000000 0.000000 ( 0.007726)
user system total real
0.010000 0.000000 0.010000 ( 0.007594)
user system total real
0.010000 0.000000 0.010000 ( 0.007664)
user system total real
0.000000 0.000000 0.000000 ( 0.009790)
Start benchmark for search 5 times by simple match query of 50 nodes (on filesystem)
--------Matched 0 nodes of 50 objects
user system total real
0.050000 0.010000 0.060000 ( 0.055999)
--------Matched 0 nodes of 50 objects
user system total real
0.110000 0.020000 0.130000 ( 0.119940)
--------Matched 0 nodes of 50 objects
user system total real
0.050000 0.010000 0.060000 ( 0.056107)
--------Matched 0 nodes of 50 objects
user system total real
0.050000 0.010000 0.060000 ( 0.067170)
--------Matched 0 nodes of 50 objects
user system total real
0.060000 0.010000 0.070000 ( 0.073780)
Start benchmark for search 5 times with wildcard match query of 50 nodes (on filesystem)
--------Matched 8 nodes of 50 objects
user system total real
0.040000 0.010000 0.050000 ( 0.052666)
--------Matched 7 nodes of 50 objects
user system total real
0.050000 0.010000 0.060000 ( 0.055499)
--------Matched 7 nodes of 50 objects
user system total real
0.040000 0.010000 0.050000 ( 0.052032)
--------Matched 4 nodes of 50 objects
user system total real
0.160000 0.010000 0.170000 ( 0.165832)
--------Matched 6 nodes of 50 objects
user system total real
0.040000 0.010000 0.050000 ( 0.052942)
Done! Data will be cleared.
============================

Preparing fake node objects upto 100...
user system total real
1.320000 0.160000 1.480000 ( 1.497286)
Done!
============================
Start benchmark for pick up one node 5 times of 100 nodes (on filesystem)
user system total real
0.010000 0.010000 0.020000 ( 0.008849)
user system total real
0.010000 0.000000 0.010000 ( 0.008563)
user system total real
0.000000 0.000000 0.000000 ( 0.008509)
user system total real
0.010000 0.000000 0.010000 ( 0.008801)
user system total real
0.010000 0.000000 0.010000 ( 0.011175)
Start benchmark for search 5 times by simple match query of 100 nodes (on filesystem)
--------Matched 0 nodes of 100 objects
user system total real
0.220000 0.030000 0.250000 ( 0.246197)
--------Matched 0 nodes of 100 objects
user system total real
0.130000 0.030000 0.160000 ( 0.163799)
--------Matched 0 nodes of 100 objects
user system total real
0.260000 0.030000 0.290000 ( 0.284852)
--------Matched 0 nodes of 100 objects
user system total real
0.140000 0.030000 0.170000 ( 0.171276)
--------Matched 0 nodes of 100 objects
user system total real
0.250000 0.030000 0.280000 ( 0.276693)
Start benchmark for search 5 times with wildcard match query of 100 nodes (on filesystem)
--------Matched 15 nodes of 100 objects
user system total real
0.150000 0.030000 0.180000 ( 0.188616)
--------Matched 19 nodes of 100 objects
user system total real
0.260000 0.030000 0.290000 ( 0.291682)
--------Matched 19 nodes of 100 objects
user system total real
0.150000 0.040000 0.190000 ( 0.189395)
--------Matched 19 nodes of 100 objects
user system total real
0.330000 0.040000 0.370000 ( 0.380891)
--------Matched 21 nodes of 100 objects
user system total real
0.150000 0.030000 0.180000 ( 0.173191)
Done! Data will be cleared.
============================

Preparing fake node objects upto 500...
user system total real
6.940000 0.900000 7.840000 ( 7.878502)
Done!
============================
Start benchmark for pick up one node 5 times of 500 nodes (on filesystem)
user system total real
0.020000 0.000000 0.020000 ( 0.014835)
user system total real
0.010000 0.000000 0.010000 ( 0.015202)
user system total real
0.080000 0.010000 0.090000 ( 0.086860)
user system total real
0.010000 0.000000 0.010000 ( 0.015377)
user system total real
0.020000 0.000000 0.020000 ( 0.015287)
Start benchmark for search 5 times by simple match query of 500 nodes (on filesystem)
--------Matched 0 nodes of 500 objects
user system total real
4.500000 0.810000 5.310000 ( 5.316760)
--------Matched 0 nodes of 500 objects
user system total real
4.530000 0.790000 5.320000 ( 5.314164)
--------Matched 0 nodes of 500 objects
user system total real
4.470000 0.780000 5.250000 ( 5.270624)
--------Matched 0 nodes of 500 objects
user system total real
4.440000 0.780000 5.220000 ( 5.215047)
--------Matched 0 nodes of 500 objects
user system total real
4.830000 0.830000 5.660000 ( 5.705454)
Start benchmark for search 5 times with wildcard match query of 500 nodes (on filesystem)
--------Matched 101 nodes of 500 objects
user system total real
4.410000 0.760000 5.170000 ( 5.172413)
--------Matched 75 nodes of 500 objects
user system total real
4.540000 0.760000 5.300000 ( 5.332210)
--------Matched 74 nodes of 500 objects
user system total real
4.500000 0.760000 5.260000 ( 5.259890)
--------Matched 89 nodes of 500 objects
user system total real
4.430000 0.760000 5.190000 ( 5.200891)
--------Matched 101 nodes of 500 objects
user system total real
4.520000 0.790000 5.310000 ( 5.326836)
Done! Data will be cleared.
============================

Preparing fake node objects upto 1000...
user system total real
14.400000 1.850000 16.250000 ( 16.261022)
Done!
============================
Start benchmark for pick up one node 5 times of 1000 nodes (on filesystem)
user system total real
0.020000 0.000000 0.020000 ( 0.022240)
user system total real
0.020000 0.010000 0.030000 ( 0.023544)
user system total real
0.020000 0.000000 0.020000 ( 0.025304)
user system total real
0.020000 0.010000 0.030000 ( 0.025084)
user system total real
0.020000 0.000000 0.020000 ( 0.027034)
Start benchmark for search 5 times by simple match query of 1000 nodes (on filesystem)
--------Matched 0 nodes of 1000 objects
user system total real
18.240000 3.240000 21.480000 ( 21.577591)
--------Matched 0 nodes of 1000 objects
user system total real
17.410000 3.020000 20.430000 ( 20.477257)
--------Matched 0 nodes of 1000 objects
user system total real
17.430000 3.100000 20.530000 ( 20.582965)
--------Matched 0 nodes of 1000 objects
user system total real
17.150000 3.000000 20.150000 ( 20.186336)
--------Matched 0 nodes of 1000 objects
user system total real
17.700000 3.100000 20.800000 ( 20.857837)
Start benchmark for search 5 times with wildcard match query of 1000 nodes (on filesystem)
--------Matched 171 nodes of 1000 objects
user system total real
17.660000 3.070000 20.730000 ( 20.761662)
--------Matched 171 nodes of 1000 objects
user system total real
17.200000 3.030000 20.230000 ( 20.273779)
--------Matched 174 nodes of 1000 objects
user system total real
17.210000 3.000000 20.210000 ( 20.221381)
--------Matched 159 nodes of 1000 objects
user system total real
17.470000 3.030000 20.500000 ( 20.542159)
--------Matched 171 nodes of 1000 objects
user system total real
17.400000 3.020000 20.420000 ( 20.460623)
Done! Data will be cleared.
============================

Preparing fake node objects upto 5000...
user system total real
69.430000 9.860000 79.290000 ( 79.626666)
Done!
============================
Start benchmark for pick up one node 5 times of 5000 nodes (on filesystem)
user system total real
0.080000 0.010000 0.090000 ( 0.093334)
user system total real
0.070000 0.020000 0.090000 ( 0.091920)
user system total real
0.080000 0.020000 0.100000 ( 0.090885)
user system total real
0.180000 0.010000 0.190000 ( 0.196649)
user system total real
0.070000 0.020000 0.090000 ( 0.097520)
Start benchmark for search 5 times by simple match query of 5000 nodes (on filesystem)
E, [2014-08-26T20:23:35.452344 #31513] ERROR -- : Actor crashed!
Ridley::Errors::TimeoutError: too many connection resets (due to Net::ReadTimeout - Net::ReadTimeout) after 0 requests on 2597410980, last used 125.772078 seconds ago