1Apache::TestSmoke(3) User Contributed Perl Documentation Apache::TestSmoke(3)
2
6 Apache::TestSmoke - Special Tests Sequence Failure Finder
7
9 # get the usage and the default values
10 % t/SMOKE -help
11 # repeat all tests 5 times and try 20 random iterations
13 # and save the report into the file 'myreport'
14 % t/SMOKE -times=5 -iterations=20 -report=myreport
15 # run all tests default number of iterations, and repeat tests
17 # default number of times
18 % t/SMOKE
19 # same as above but work only the specified tests
21 % t/SMOKE foo/bar foo/tar
22 # run once a sequence of tests in a non-random mode
24 # e.g. when trying to reduce a known long sequence that fails
25 % t/SMOKE -order=rotate -times=1 foo/bar foo/tar
26 # show me each currently running test
28 # it's not the same as running the tests in the verbose mode
29 % t/SMOKE -verbose
30 # run t/TEST, but show any problems after *each* tests is run
32 # useful for bug reports (it actually runs t/TEST -start, then
33 # t/TEST -run for each test separately and finally t/TEST -stop
34 % t/SMOKE -bug_mode
35 # now read the created report file
37
39 The Problem
40 When we try to test a stateless machine (i.e. all tests are indepen‐
42 dent), running all tests once ensures that all tested things properly
43 work. However when a state machine is tested (i.e. where a run of one
44 test may influence another test) it's not enough to run all the tests
45 once to know that the tested features actually work. It's quite possi‐
46 ble that if the same tests are run in a different order and/or repeated
47 a few times, some tests may fail. This usually happens when some tests
48 don't restore the system under test to its pristine state at the end of
49 the run, which may influence other tests which rely on the fact that
50 they start on pristine state, when in fact it's not true anymore. In
51 fact it's possible that a single test may fail when run twice or three
52 times in a sequence.
53 The Solution
55 To reduce the possibility of such dependency errors, it's helpful to
57 run random testing repeated many times with many different srand seeds.
58 Of course if no failures get spotted that doesn't mean that there are
59 no tests inter-dependencies, which may cause a failure in production.
60 But random testing definitely helps to spot many problems and can give
61 better test coverage.
62 Resolving Sequence Problems
64 When this kind of testing is used and a failure is detected there are
66 two problems:
67 1 First is to be able to reproduce the problem so if we think we
69 fixed it, we could verify the fix. This one is easy, just remember
70 the sequence of tests run till the failed test and rerun the same
71 sequence once again after the problem has been fixed.
72 2 Second is to be able to understand the cause of the problem. If
74 during the random test the failure has happened after running 400
75 tests, how can we possibly know which previously running tests has
76 caused to the failure of the test 401. Chances are that most of the
77 tests were clean and don't have inter-dependency problem. Therefore
78 it'd be very helpful if we could reduce the long sequence to a min‐
79 imum. Preferably 1 or 2 tests. That's when we can try to understand
80 the cause of the detected problem.
81 This utility attempts to solve both problems, and at the end of each
83 iteration print a minimal sequence of tests causing to a failure. This
84 doesn't always succeed, but works in many cases.
85 This utility:
87 1 Runs the tests randomly until the first failure is detected. Or
89 non-randomly if the option -order is set to repeat or rotate.
90 2 Then it tries to reduce that sequence of tests to a minimum, and
92 this sequence still causes to the same failure.
93 3 (XXX: todo): then it reruns the minimal sequence in the verbose
95 mode and saves the output.
96 4 It reports all the successful reductions as it goes to STDOUT and
98 report file of the format: smoke-report-<date>.txt.
99 In addition the systems build parameters are logged into the report
101 file, so the detected problems could be reproduced.
102 5 Goto 1 and run again using a new random seed, which potentially
104 should detect different failures.
105
107 Currently for each reduction path, the following reduction algorithms
108 get applied:
109 1 Binary search: first try the upper half then the lower.
111 2 Random window: randomize the left item, then the right item and
113 return the items between these two points.
114
116 t/SMOKE.PL is driving this module, if you don't have it, create it:
117 #!perl
119 use strict;
121 use warnings FATAL => 'all';
122 use FindBin;
124 use lib "$FindBin::Bin/../Apache-Test/lib";
125 use lib "$FindBin::Bin/../lib";
126 use Apache::TestSmoke ();
128 Apache::TestSmoke->new(@ARGV)->run;
130 usually Makefile.PL converts it into t/SMOKE while adjusting the perl
132 path, but you create t/SMOKE in first place as well.
133
135 Stas Bekman
136perl v5.8.8 2006-11-19 Apache::TestSmoke(3)