NAME
Test::System - Test suite oriented for testing system administration tasks
SYNOPSIS
use Test::System;
my $suite = Test::System->new(
format => 'consoletable',
nodes => 'example.com',
test_groups => '~/code/my/system/tests.yaml'
);
$suite->runtests;
DESCRIPTION
Loads and runs a set of tests for verifying a system administration task before a possible incident happens or after a task is executed.
Tests can be run as perl scripts and actually in any format TAP::Harness supports.
Some examples of how Test::System can help you:
Before a deployment probably you want to make sure your systems and configuration files are ready to go. Also you probably want to make sure that if you needed to apply patches to your configuration files, they really got applied (even in a development server or verify directly on production servers). So here Test::System can help you verifying this in the form of a test case. Of course, you write the test case :-)
What about if you have many servers and although you have nagios to verify the speed of the CPU sometimes you need to verify it "on-demand" (perhaps when you get new servers in your datacenter). Here Test::System can help you to verify that these servers are ready to go before even they get connected to nagios or your favorite monitoring tool.
One of the things you need to keep in mind is that Test::System is NOT a monitoring system. It will let you verify things on demand, things that are repetitive and perhaps you want them to be automated.
Attributes
The module offers a list of attributes, some of them are read-only.
- test_groups
-
YAML filename that has a list of available tests. This is not required but can be useful if you want to group tests (like hardware.yaml, net.yaml, etc). Comes also handy when user does not provide a list of tests to execute, so all the tests listed in this file are executed.
An example of a YAML file:
ping: description: Test the ping and do foo and make bar code: test/foo.pl cpu: description: Test the CPU of nodes code: test/cpu.pl
- available_tests
-
Is a read only attribute that contains a list of all available tests found in the YAML file provided by
test_groups
. - nodes
-
Is an attribute that can be represented as a string (like a hostname) or as a list (where each item will be a node/hostname). This attribute has write access and is where the tests are going to be executed to.
Another way of setting this value is when
runtests()
is called, a test plan (in the form of a YAML file) can be provided and it can contain the list of nodes to use. - format
-
A write access string that has the format of how the tests should be presented, please refer to the modules available under Test::System::Output or in your custom factory (
custom_factory
attribute) class if this is the case. - available_formats
-
A list of available formats, read only.
- custom_factory
-
If you want to use your own Factory for creating your output you can set this to your class name (NOT the object).
- harness
-
Is (or will be) the harness instance once
runtests
is executed. - parameters
-
An attribute with write access permission. This attribute will transform all the items of this hash to environment variables.
The use of this attribute is very handy if you want to provide some additional data for your tests and since the tests are run in separate forks with Test::Harness then the only possible way to keep them is to make them available through the environment (
%ENV
hash).Please be warned that only scalars are stored in environment variables, those that are an array will be converted to CSV values while the rest of the data types will be lost.
In your tests if you want to use any of these parameters they will be available through the environment variables with a prefix of:
TEST_SYSTEM_
or you can use Test::System::Helper to get their values. - results
-
A hash reference that contains the results of the tests. This information is generated by the TAP::Parser::Aggregator.
passed: Number of passed tests
total: Total number of tests (passed or not)
skipped: Tests that required a SKIP
- status
-
A string (read-only) that contains a word that describes the status of all the tests. This value is also generated by TAP::Parser::Aggregator.
Methods
runtests(@tests,%options)
It will run the given test cases (@tests
). The @tests
can be:
An array of test files to execute.
Or, a string pointing to a test plan file (YAML).
If no @tests
are given but we have a test_groups
then ALL the tests listed inside this file will be executed.
The %options
is a hash of options that will be passed to the TAP::Harness object, some useful parameters are:
verbosity
By default we mute everything with
-9
.color
If you want the output (in console) to have color
formatter
Although we use Test::System::Output::Factory to offer a set of formatters you can provide your own formatter object. See
available_formats
.jobs
If you have many tests you probably want to increment this value (that defaults to
1
) so other tests can be run at the same time.
setup_parser($parser)
This method should never be called directly since this is triggered by TAP::Harness when every TAP::Parser object gets created.
This method is useful for setting the callbacks we want the parser to trigger.
$parser
should be a TAP::Parser object.
pretests_verification(@tests, %options)
Does some verification before the tests are executed. This method gets called by runtests
.
The parameters it accepts are the same parameters passed to runtests
with the main difference the tests are already filtered (eg, if a test plan in YAML was provided we will use it or we will use all the tests listed inside the test_groups
.
what_happened($parser, $result_test)
Similar to setup_parser, this should never be called since it will be triggered by TAP::Parser after each (sub)tests gets executed.
By default we check the reason of why the test failed so later we provide a simple hash to parse with all the tests and the reasons of why they failed (if this could be the case). And because we need to know what TAP::Parser instance we are processing we need to ask for it as a first parameter.
Parameters:
$parser
should be a TAP::Parser$result_test
should be a TAP::Parser::Result::Test.
set_results($aggregator)
This is a callback and is only called/triggered by TAP::Harness when all the tests are done. It will fill the results
attribute with information of everything.
Refer to the documentation of results for more information.
The paramenter $aggregator
is a TAP::Parser::Aggregator.
prepare_environment()
Prepares the environment by settings the needed environment values so they can be used later by the tests
clean_environment()
Cleans/deletes all the environment variables that match TEST_SYSTEM_*
get_tests_from_test_plan($yaml_file, $do_not_fill_parameters, $do_not_fill_nodes)
Reads the given tests yaml file ($yaml_file
). This YAML file should have at least a list of tests (in the form of a hash) and optionally can also have parameters the tests should contain.
Although this method is used mostly internally there's the option to call it as any other method Test::System offers.
By default it will fill the parameters of your Test::System instance but by passing $do_not_fill_parameters
(second parameter) as true or something that Perl understands as true then it will skip the part. This should be presented as a hash in YAML syntax.
The above apply also to $do_not_fill_nodes
(third parameter). This should be presented as an array in YAML syntax.
Once the file is read an array with all the tests will be returned.
An example of a YAML test plan file can be found inside the examples
directory or:
tests:
- ping
- cpu
- memory
parameters:
foo: bar
bar: zoo
nodes:
- pablo.com.mx
- example.com
SEE ALSO
Take a look to an awesome and pretty similar CPAN module: Test::Server.
AUTHOR
Pablo Fischer (pablo@pablo.com.mx).
COPYRIGHT
Copyright (C) 2009 by Pablo Fischer.
This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.