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
|
Blobs Server Multiprocessing Improvement Assessment
===================================================
The code in this directory has the purpose of showing whether multiprocessing
improves the Soledad Blobs Server or not.
It uses a haproxy instance listening on local port 8000 to route between
different processes listening on ports 8001-8008. The number of processes
depends on the testing scenario.
Then, it runs a set of tests using 1, 2, 4 and 8 blobs server processes and
stores the results in a text file. Those results can then be graphed and
compared visually.
Dependencies
------------
To run the testing script, you will need to have:
- A python virtualenv with Soledad Server installed on it.
- A working haproxy installation.
- curl
To generate the graph, you will need:
- numpy
- matplotlib/pyplot
Description of files
--------------------
.
├── blobs-in-parallel.png - A graph generated from ./results.txt
├── blobs-server.py - A simple blobs server.
├── graph.py - The graphing script.
├── haproxy - A directory with haproxy config.
├── makefile - Rules for starting up pieces of the test.
├── multiproc.py - A service that spawns multiple blobs servers.
├── proxies - Study with different possibilities of proxy.
├── README.txt - This file.
├── request.py - A stressing blobs client.
├── results - A directory with some results stored.
├── results.txt - The file that ./graph.py reads from.
└── run-test.sh - The script to run tests and produce output.
Actions
-------
The following set of actions are tested in each scenario, and compose the
X axis of the final graph:
- baseline: a simple GET / to the root of the webserver, returning an empty
page. Nothing can be faster than that.
- list: a simple GET /blobs/some-user which lists the current blobs. In the
present case, the result will always be an empty set of blobs.
- put: a PUT /blobs/some-user/some-blob-id. The size of the blob depends on
the scenarios configured in ./run-test.sh.
- get: a PUT /blobs/some-user/some-blob-id followed by a GET to the same
blob.
- flag: a PUT /blobs/some-user/some-blob-id followed by a POST to the same
blob setting one flag.
- delete: a PUT /blobs/some-user/some-blob-id followed by a DELETE of the
same blob.
When testing, each action is run 5 times (so the numbers in ./results.txt are
the accumulated time of the 5 runs) and the mean of the time taken is used when
graphing.
Usage
-----
Configure the scenarios you want tests to be run by editing ./run-test.sh, and
then run the script. It will:
- iterate through desired number of processes and startup haproxy and
multiproc blobs server for each scenario.
- iterate through all the actions tested (baseline, put, put+get, etc).
- iterate through amount x size blobs scenarios.
- delete the contents of /tmp/blobs/* and run tests for each scenario.
Check the makefile for rules to help debugging (i.e. `make server` and `make
get`).
|
