NAME
Schedule::Pluggable - Flexible Perl Process Scheduler
SYNOPSIS
EXAMPLE #1: Simple Run in Series
use Schedule::Pluggable;
my $p = Schedule::Pluggable->new;
my $status = $p->run_in_series( [ qw/command1 command2 command3/ ] );
EXAMPLE #2: Simple Run in Parallel
use Schedule::Pluggable;
my $p = Schedule::Pluggable->new;
my $status = $p->run_in_parallel( [ qw/command1 command2 command3/ ] );
EXAMPLE #3: With Job Names
use Schedule::Pluggable;
my $p = Schedule::Pluggable->new;
my @jobs = (
{ name => "FirstJob", command => "somescript.sh" },
{ name => "2nd Job", command => sub { do_something; } },
);
my $status = $p->run_schedule( \@jobs );
EXAMPLE #4: With Prerequsites
use Schedule::Pluggable;
my $p = Schedule::Pluggable->new;
my $jobs = [ { name => "FirstJob",
command => "somescript.sh" },
{ name => "SecondJob",
command => sub { do_something; },
prerequisites => [qw/FirstJob/] },
];
my $status = $p->run_schedule( $jobs );
EXAMPLE #5: Same as #4 but with dependencies
use Schedule::Pluggable;
my $p = Schedule::Pluggable->new;
my $jobs = [ { name => "FirstJob",
command => "somescript.sh",
dependencies => [qw/SecondJob/] },
{ name => "SecondJob",
command => sub { do_something; } },
];
my $status = $ps->run_schedule( $jobs );
EXAMPLE #5: With Groups
use Schedule::Pluggable;
my $p = Schedule::Pluggable->new;
my $jobs = ( { name => "one", command => "one.sh", dependencies => [ qw/Reports/ ] },
{ name => "two", command => "two.pl" }, groups => [ qw/Reports/] },
{ name => "three", command => "three.pl" }, groups => [ qw/Reports/] },
{ name => "four", command => "four.ksh" }, prerequisites => [ qw/Reports/] },
);
my $status = $p->run_schedule( $jobs );
EXAMPLE #6: Getting the config from an XML file
use Schedule::Pluggable (JobsConfig => 'JobsFromXML');
my $p = Schedule::Pluggable->new;
my $status = $p->run_schedule({XMLFile => 'path to xml file'});
XMlFile in following format :-
<?xml version="1.0"?>
<Jobs>
<Job name='Job1' command='succeed.pl'>
<params>3</params>
<dependencies>second</dependencies>
</Job>
<Job name='Job2' command='fail.pl'>
<params>3</params>
<group>second</group>
</Job>
...
<Jobs>DESCRIPTION
Schedule::Pluggable is a perl module which provides a simple but powerful way of running processes in a controlled way. In true perl fashion it makes simple things easy and complicated things possible. It also uses a system of plugins so you can change it's behaviour to suit your requirements by supplying your own plugins. For most cases the default plugins will suffice however.
OPTIONS
You can override the default behaviour of Schedule::Pluggable by supplying options with the use statement in for form of a hash
i.e.
use Schedule::Pluggable ( Option => "value' );
The Following options are supported :-
JobsConfig
Specifies which plugin to use to provide the job configuration - defaults to JobsFromData which expects you to supply the job configuration in an array
Each plugin is expected to
e.g.
use Schedule::Pluggable ( JobsConfig => 'JobsFromSomeWhere' );
Currently the available values are :-
- JobsFromArray
-
The default which activates the role 'Schedule::Pluggable::Plugin::JobsFromData' which as the name suggests expects the job configuration to be be supplied as an reference to an array of jobs to run.
- JobsFromXML
-
Activates plugin Schedule::Pluggable::Plugin::JobsFromXML which obtains the jobs configuration from an XML file
This enables you to specify a different source for the config by supplying an appropriate plugin for it - see writing Plugins for details
EventHandler
Controls what happens when an event happens like a jobs starting a job failing e.t.c Defaults to DefaultEventHandler which is a plugin Schedule::Pluggable::Plugin::DefaultEventHandler Here is what is passed depending on event type
- JobName - Always passed
- Command - Always passed
- Stdout - Passed on JobStdout and JobSucceeded only
- Stderr - Passed on JobStderr and JobFailed
- ReturnValue - Passed on JobFailed
This handler uses other configuration options to control it's behaviour as follows :-
EventsToReport
Comma separated list of events to report on or 'all' for al of them of 'none' for none of them Defaults to qq/JobFailed,JobSucceeded,JobStderr/
e.g.
use Schedule::Pluggable ( EventsToReport => qw/JobQueued,JobFailed,JobSucceeded,JobStderr/ );
PrefixWithTimeStamp
whether to prefix messages with the current time in dd/mm/yyyy HH::MM::SS format.
Defaults to 1 (timestamp is produced)
MessagesTo
where messages are sent - stdout by default
If supplied a filehandle, will call the print method on it and pass the details, for anything else will call directly. So this could be a Log::Log4perl method e.g. $log->info or $log->{ Category }->info
ErrorsTo
where error messages are sent - stderr by default
If supplied a filehandle, will call the print method on it and pass the details, for anything else will call directly. So this could be a Log::Log4perl method e.g. $log->error or $log->{ Category }->error
e.g.
use Schedule::Pluggable ( ErrorsTo => \&my_logger );
or
use Schedule::Pluggable; my $p = Schedule::Pluggable->new( MessagesTo => \&my_logger );
JOB CONIFIGURATION FORMAT
A Job entry can be a scalar value in which case it is assumed to contain a command to run or a hash containing some or all of the following :-
- name
-
the name of the job
- command - command to run
- params - array of parameters to the command
- groups - array of groups to which the job belongs
- prerequisites - array of jobs or groups which must have completed successfully before job with start
- dependencies - array of jobs or groups which must wait until this job has completed successfully before they will start
Obviously the bare minimum is to supply a command to run If a name is not supplied, it will be allocated one in the format Jobn where n is an incrementing number starting at 1 and increases with each job specified
METHODS
- run_in_series ( $job_specification )
-
Utility method to run the supplied jobs in series by creating dependencies where each job is dependant on the previous one and then calls run_schedule with the revised definition
- run_in_parallel ( $job_specification )
-
Runs the supplied jobs in parallel
Utility method to run the supplied jobs in parallel by removing and dependencies which are defined and the call run_schedule
- run_schedule ( $job_specification )
-
The main method of the module - takes a supplied job definition - processes the information to validate and expand the definition and then runs the jobs as specified. When any event occurs, the appropriate callback is called if required to report on progress and on completion returns a structure detailing what happened in the following format :-
$status = { TotalJobs => <total number of jobs in schedule>, TotalFailed => <number of jobs which failed>, TotalFinished => <number of jobs which finished>, TotalSucceeded' => <number of successfull jobs>' LastUpdate => 'dd/mm/yyyy hh::mm::ss', Failed => { <Job which failed> => { status => <return value of job>, stderr => [ 'error line 1', .... ], }, }, Jobs => { <Job Name> => { name => <Job Name>, command => <command> status => <return value of command> Pid => <Process Id>, timestarted => 'dd/mm/yyyy hh::mm::ss', timefinished=> 'dd/mm/yyyy hh::mm::ss', stderr => [ 'error line 1', .... ], stdout => [ 'output line 1', .... ], }, ........ }, };
- BUILDARGS Handles module options via import or passed on objet creation
- BUILD Handles loading plugins
JOB DEFINITIONS
Jobs are specified as reference to an array which can contain either a list of commands ot run or as hash values
- scalar values containing commands to run
- hashes containing at least one key 'command' with the value containing the command to run
AUTHOR -
Tony Edwardson <Tony@Edwardson.co.uk>
KNOWN ERRORS
None yet - let me know if you find any