++ed by:
4 non-PAUSE users
Author image Alan Kasindorf


Gearman::Task - a task in Gearman, from the point of view of a client


    my $task = Gearman::Task->new("add", "1+2", {




Gearman::Task is a Gearman::Client's representation of a task to be done.


Gearman::Task->new($func, $arg, \%options)

Creates a new Gearman::Task object, and returns the object.

$func is the function name to be run. (that you have a worker registered to process)

$arg is an opaque scalar or scalarref representing the argument(s) to pass to the distributed function. If you want to pass multiple arguments, you must encode them somehow into this one. That's up to you and your worker.

%options can contain:

  • uniq

    A key which indicates to the server that other tasks with the same function name and key will be merged into one. That is, the task will be run just once, but all the listeners waiting on that job will get the response multiplexed back to them.

    Uniq may also contain the magic value "-" (a single hyphen) which means the uniq key is the contents of the args.

  • on_complete

    A subroutine reference to be invoked when the task is completed. The subroutine will be passed a reference to the return value from the worker process.

  • on_fail

    A subroutine reference to be invoked when the task fails (or fails for the last time, if retries were specified). No arguments are passed to this callback. This callback won't be called after a failure if more retries are still possible.

  • on_retry

    A subroutine reference to be invoked when the task fails, but is about to be retried.

    Is passed one argument, what retry attempt number this is. (starts with 1)

  • on_status

    A subroutine reference to be invoked if the task emits status updates. Arguments passed to the subref are ($numerator, $denominator), where those are left up to the client and job to determine.

  • retry_count

    Number of times job will be retried if there are failures. Defaults to 0.

  • high_priority

    Boolean, whether this job should take priority over other jobs already enqueued.

  • timeout

    Automatically fail, calling your on_fail callback, after this many seconds have elapsed without an on_fail or on_complete being called. Defaults to 0, which means never. Bypasses any retry_count remaining.

  • try_timeout

    Automatically fail, calling your on_retry callback (or on_fail if out of retries), after this many seconds have elapsed. Defaults to 0, which means never.


Returns bool: whether or not task is totally done (on_failure or on_complete callback has been called)