[REPORT] Performance Test Samba vs Windows
joeg at catalyst.net.nz
joeg at catalyst.net.nz
Mon May 14 03:38:07 UTC 2018
It seems like some people have trouble to receive my original email
because of the images.
This is the text version:
ubuntu at wpts-s1-traffic-runner:~/samba$ python summary.py
r\S 1 2 3 4 5 6
7 8 9 10 11 12
1 0 0 0 0 0 0
0 0 0 0 0 0
2 0 0 0 0 0 0
0 0 0 0 0 0
3 0 0 0 0 0 0
0 0 0 0 0 0
4 0 0 0 41.119 56.732 66.447 72.572
76.956 95.436 102.396 110.454 0
5 0 0 0 49.145 70.014 74.849 84.013
89.707 107.108 116.287 103.106 0
6 0 0 0 68.731 78.698 93.159 95.705
115.973 114.349 112.678 130.282 0
7 0 0 0 71.490 85.902 108.542 108.586
113.590 126.788 119.140 101.968 0
8 0 0 0 84.228 103.297 112.420 108.141
112.146 125.492 109.455 101.155 0
9 0 0 0 88.669 111.279 121.480 121.692
138.248 115.464 131.850 98.632 0
10 0 0 0 107.833 115.134 126.926 132.568
116.088 136.371 97.090 126.787 0
11 0 0 0 114.108 111.883 126.213 139.040
141.352 106.134 124.904 118.583 0
12 0 0 0 109.988 120.804 142.387 133.999
124.241 131.333 107.382 111.447 0
13 0 0 0 133.787 115.635 135.760 144.311
113.029 109.257 124.848 135.680 0
14 0 0 0 142.868 118.900 139.201 131.333
132.071 113.330 116.351 118.728 0
15 0 0 0 124.355 132.024 146.237 130.092
107.181 125.418 121.196 126.104 0
16 0 0 0 123.211 140.214 135.858 109.205
146.865 137.027 111.136 121.051 0
17 0 0 0 120.864 149.945 116.799 147.768
119.295 126.733 119.915 107.164 0
18 0 0 0 140.311 138.099 108.622 126.043
127.352 143.128 134.762 93.017 0
19 0 0 0 145.346 133.715 120.586 139.703
120.015 118.425 83.349 106.214 0
20 0 0 0 144.652 130.634 110.613 123.414
151.610 135.753 93.042 85.515 0
21 0 0 0 131.350 122.082 121.926 146.542
132.719 126.720 120.016 91.363 0
22 0 0 0 140.259 146.325 129.648 142.637
142.712 132.569 124.193 84.189 0
23 0 0 0 *169.764 139.553 91.046 137.663
159.753 138.276 102.191 110.595 0
24 0 0 0 148.510 142.059 141.010 137.082
98.231 96.259 109.009 69.462 0
25 0 0 0 0 0 0
0 0 0 0 0 0
max: 169.764
ubuntu at wpts-s1-traffic-runner-win1:~/samba$ python summary.py
r\S 1 2 3 4 5 6
7 8 9 10 11 12
1 0 0 0 0 0 0
0 0 0 0 0 0
2 0 0 0 0 0 0
0 0 0 0 0 0
3 0 0 0 0 0 0
0 0 0 0 0 0
4 0 0 0 37.499 51.081 47.946 78.658
73.416 92.505 95.842 100.820 0
5 0 0 0 55.603 57.470 74.442 86.191
113.089 111.861 115.396 142.818 0
6 0 0 0 58.732 85.243 92.833 102.397
108.455 136.802 134.110 136.791 0
7 0 0 0 66.669 91.372 100.347 132.541
159.121 137.524 166.432 131.660 0
8 0 0 0 76.959 113.768 130.787 148.635
140.318 155.053 185.398 172.079 0
9 0 0 0 104.442 124.219 135.548 156.073
159.233 161.505 165.132 111.609 0
10 0 0 0 121.444 128.814 170.306 172.741
173.592 155.078 137.541 160.373 0
11 0 0 0 119.806 147.530 171.919 175.189
156.816 189.379 176.344 155.873 0
12 0 0 0 103.682 145.108 174.202 172.044
206.165 129.995 157.405 150.218 0
13 0 0 0 151.011 151.274 172.072 189.949
161.779 176.042 137.990 140.948 0
14 0 0 0 146.589 166.755 213.144 189.248
166.505 145.031 138.344 150.909 0
15 0 0 0 149.442 190.469 200.051 160.482
143.324 151.437 141.908 130.576 0
16 0 0 0 173.891 202.137 204.066 172.500
176.371 166.698 127.024 141.724 0
17 0 0 0 170.829 180.184 196.766 181.791
147.663 155.474 99.551 127.230 0
18 0 0 0 188.009 191.858 155.009 172.549
147.957 152.508 124.296 129.689 0
19 0 0 0 201.694 225.261 190.230 157.619
152.969 136.545 130.671 125.775 0
20 0 0 0 203.151 209.354 165.844 166.927
141.296 140.057 139.265 125.006 0
21 0 0 0 212.496 217.526 171.679 168.568
168.101 145.787 111.281 106.252 0
22 0 0 0 184.216 *235.421 180.842 156.551
158.798 149.434 115.148 113.642 0
23 0 0 0 201.699 210.266 191.049 153.199
153.461 110.562 110.072 112.646 0
24 0 0 0 222.205 150.690 155.598 176.060
137.402 123.665 115.973 106.362 0
25 0 0 0 0 0 0
0 0 0 0 0 0
max: 235.421
And I should mention that Samba is running in prefork mode in this test:
sudo bin/samba -M prefork --option='prefork children = 4'
On 14/05/18 15:16, Garming Sam via samba-technical wrote:
> There's some interesting observations that you can make about the
> overall traffic, without looking at too much detail of the actual runs.
>
> 1) In terms of maximum traffic (which in our examples has been quite
> LDAP based), Windows is at most 50% faster. This number likely shrinks
> depending on whether or not the service levels (e.g. 99% success) at
> such high load are acceptable or not.
>
> 2) Looking at median packet response times, we generally spend twice as
> long on processing a packet. However, the difference between our 95
> percentile values is often much smaller than this and some of the mean
> values are actually at parity.
>
> 3) With small to moderate loads, Samba handles a similar amount of
> operations per second to Windows. For the values shown, it actually
> seems to get more throughput (although not incredibly so) for a good
> chunk of the table shown.
>
> It's actually kind of surprising that Samba doesn't perform all that
> much worse than Windows does. It's certainly not off by an order of
> magnitude (or two), although this testing has been run in the prefork
> mode. Our previous analyses showed much smaller numbers and that's
> likely due to the limitations of being stuck with a single-process.
>
>
> Cheers,
>
> Garming
>
> On 14/05/18 14:45, joeg at catalyst.net.nz wrote:
>> Hi Team,
>>
>> I've been trying to make the Samba performance test tool work against
>> Windows.
>>
>> After a lot of effort, Garming and I fixed all the errors for
>> different packets, and finally it's working(fixes are merged into master).
>>
>> Base on that, we can send traffic to Samba and Windows, get the
>> maximum load, and compare the performance.
>>
>> An example of the command I was using:
>>
>>
>> script/traffic_replay -U Administrator%PASSWORD --realm
>> krb.samba.site --workgroup KRB --fixed-password FIXED-PASSWORD -r 4 -S
>> 4 traffic-sample-1-model.txt dc1.krb.samba.site 2>&1 >
>> traffic_replay_stats/dc1_r4_S4.txt
>>
>>
>> By changing the combination of different -r and -S, we can get test
>> results and save to files, then parse files to get summary tables as
>> below.
>>
>> The number displayed is "*Successful operations per s**econd*", the 0
>> ones are combinations we think not important and skipped.
>>
>>
>> Test Result for Samba (master code on 2018.05.11):
>>
>>
>>
>> Test Result for Windows Server 2012r2:
>>
>>
>>
>>
>> I am running the test with 4 servers in Catalyst Cloud which is Openstack:
>>
>>
>>
>> All data are attached, hope this can help the team to understand how
>> is Samba performing with compare to Windows.
>> --
>> Joe Guo
>> joeg at catalyst.net.nz
>> Catalyst IT
--
Joe Guo
joeg at catalyst.net.nz
Catalyst IT
More information about the samba-technical
mailing list