~drizzle-trunk/drizzle/development

2317.1.4 by patrick crews
Updated docs for sysbench
1
**********************************
2
sysbench
3
**********************************
4
5
6
7
Description
8
===========
9
dbqp's sysbench mode allows a user to run a specific iteration of a sysbench test (eg an oltp readonly run at concurrency = 16)
10
11
12
Requirements
13
============
14
15
The SYSBENCH command requires that the Drizzle library header be installed. Simply build Drizzle and do::
16
17
    $> sudo make install
18
19
This will install the right headers.
20
21
The SYSBENCH command also requires installation of the drizzle-sysbench program, which can be installed from source like so::
22
23
    $> bzr branch lp:~drizzle-developers/sysbench/trunk  drizzle-sysbench
24
    $> cd drizzle-sysbench
25
    $> ./autogen.sh && ./configure && make && sudo make install
26
27
Make sure sysbench is then in your path
28
29
30
sysbench / dbqp tests
31
=====================
32
33
A sysbench test defines a run for a particular concurrency.  There are suites for readonly and readwrite.
34
They are currently broken down this way as an experiment - we are open to other ways of organizing these tests::
35
36
    [test_info]
37
    comment = 16 threads
38
39
    [test_command]
40
    command = sysbench --max-time=240 --max-requests=0 --test=oltp --db-ps-mode=disable --drizzle-table-engine=innodb --oltp-read-only=on --oltp-table-size=1000000 --drizzle-mysql=on --drizzle-user=root --drizzle-db=test --drizzle-port=$MASTER_MYPORT --drizzle-host=localhost --db-driver=drizzle --num-threads=16
41
42
    [test_servers]
43
    servers = [[innodb.buffer-pool-size=256M innodb.log-file-size=64M innodb.log-buffer-size=8M innodb.thread-concurrency=0 innodb.additional-mem-pool-size=16M table-open-cache=4096 table-definition-cache=4096 mysql-protocol.max-connections=2048]]
44
45
Running tests
46
=========================
47
48
There are several different ways to run tests using :doc:`dbqp` 's sysbench mode.
49
50
It should be noted that unless :option:`--force` is used, the program will
51
stop execution upon encountering the first failing test. 
52
:option:`--force` is recommended if you are running several tests - it will
53
allow you to view all successes and failures in one run.
54
55
Running individual tests
56
------------------------
57
If one only wants to run a few, specific tests, they may do so this way::
58
59
    ./dbqp --mode=sysbench [OPTIONS] test1 [test2 ... testN]
60
61
Running all tests within a suite
62
--------------------------------
63
Many of the tests supplied with Drizzle are organized into suites.  
64
65
The tests within drizzle/tests/randgen_tests/main are considered the 'main' suite.  
66
Other suites are also subdirectories of drizzle/tests/randgen_tests.
67
68
To run the tests in a specific suite::
69
70
    ./dbqp --mode=sysbench [OPTIONS] --suite=SUITENAME
71
72
Running specific tests within a suite
73
--------------------------------------
74
To run a specific set of tests within a suite::
75
76
    ./dbqp --mode=sysbench [OPTIONS] --suite=SUITENAME TEST1 [TEST2..TESTN]
77
78
Calling tests using <suitename>.<testname> currently does not work.
79
One must specify the test suite via the :option:`--suite` option.
80
81
82
Running all available tests
83
---------------------------
84
One would currently have to name all suites, but the majority of the working tests live in the main suite
85
Other suites utilize more exotic server combinations and we are currently tweaking them to better integrate with the 
86
dbqp system.  The slave-plugin suite does currently have a good config file for setting up simple replication setups for testing.
87
To execute several suites' worth of tests::
88
89
    ./dbqp --mode=sysbench [OPTIONS] --suite=SUITE1, SUITE2, ...SUITEN
90
91
Interpreting test results
92
=========================
93
The output of the test runner is quite simple.  Every test should pass.
94
In the event of a test failure, please take the time to file a bug here:
95
*https://bugs.launchpad.net/drizzle*
96
97
During a run, the program will provide the user with:
98
  * test name (suite + name)
99
  * test status (pass/fail/skipped)
100
  * time spent executing each test
101
102
Example output::
103
104
    20110601-191706  ===============================================================
105
    20110601-191706  TEST NAME                                  [ RESULT ] TIME (ms)
106
    20110601-191706  ===============================================================
107
    20110601-191706  readonly.concurrency_16                    [ pass ]   240019
108
    20110601-191706  max_req_lat_ms: 21.44
109
    20110601-191706  rwreqps: 4208.2
110
    20110601-191706  min_req_lat_ms: 6.31
111
    20110601-191706  deadlocksps: 0.0
112
    20110601-191706  tps: 150.29
113
    20110601-191706  avg_req_lat_ms: 6.65
114
    20110601-191706  95p_req_lat_ms: 7.02
115
    20110601-191706  ===============================================================
116
    20110601-191706 INFO Test execution complete in 275 seconds
117
    20110601-191706 INFO Summary report:
118
    20110601-191706 INFO Executed 1/1 test cases, 100.00 percent
119
    20110601-191706 INFO STATUS: PASS, 1/1 test cases, 100.00 percent executed
120
    20110601-191706 INFO Spent 240 / 275 seconds on: TEST(s)
121
    20110601-191706 INFO Test execution complete
122
    20110601-191706 INFO Stopping all running servers...
123