Bio::EnsEMBL::Hive::Meadow Class Reference

Inheritance diagram for Bio::EnsEMBL::Hive::Meadow:

Public Member Functions
public	get_meadow_major_version ()
public	get_meadow_version ()
public	check_version_compatibility ()
public	new ()
protected	_init_meadow ()
public	cached_name ()
public	type ()
public	get_current_hostname ()
public	signature ()
public	pipeline_name ()
public	runWorker_path ()
public	job_name_prefix ()
public	job_array_common_name ()
public	name ()
public	get_current_worker_process_id ()
public	status_of_all_our_workers ()
public	check_worker_is_alive_and_mine ()
public	kill_worker ()
public	parse_report_source_line ()
public	get_report_entries_for_process_ids ()
public	get_report_entries_for_time_interval ()
public	submit_workers_return_meadow_pids ()
public	run_on_host ()
public	cleanup_temp_directory ()
Public Member Functions inherited from Bio::EnsEMBL::Hive::Configurable
public	config ()
public	context ()
public	config_get ()
public	config_set ()

Detailed Description

Description

    Meadow is an abstract interface to the queue manager.

    A Meadow knows how to check&change the actual status of Workers on the farm.

Definition at line 12 of file Meadow.pm.

Member Function Documentation

_init_meadow()

protected Bio::EnsEMBL::Hive::Meadow::_init_meadow

(

)

Undocumented method

Code:

cached_name()

public Bio::EnsEMBL::Hive::Meadow::cached_name

(

)

    Title   :  cached_name
    Function:  Wrapper around name() that caches its return value.
               This is because (1) it can be expensive to get the name
               (e.g. calling an external command), and (2) the name of a
               Meadow is not expected to change through the life of the
               agent.

Code:

check_version_compatibility()

public Bio::EnsEMBL::Hive::Meadow::check_version_compatibility

(

)

Undocumented method

Code:

check_worker_is_alive_and_mine()

public Bio::EnsEMBL::Hive::Meadow::check_worker_is_alive_and_mine

(

)

    Title   :  check_worker_is_alive_and_mine
    Function:  Tells whether the given worker lives in the current Meadow
               and belongs to the current user.

Code:

cleanup_temp_directory()

public Bio::EnsEMBL::Hive::Meadow::cleanup_temp_directory

(

)

    Title   :  cleanup_temp_directory
    Function:  Cleanup the temp directory assigned to the worker

Code:

get_current_hostname()

public Bio::EnsEMBL::Hive::Meadow::get_current_hostname

(

)

    Title   :  get_current_hostname
    Function:  Returns the "current" hostname (most UNIX-based Meadows will simply use this base method)

Code:

get_current_worker_process_id()

public Bio::EnsEMBL::Hive::Meadow::get_current_worker_process_id

(

)

    Title   :  get_current_worker_process_id
    Function:  Called by a worker to find its process_id. At any point in
               time, the triple (meadow_type, meadow_name, process_id)
               should be unique

Code:

get_meadow_major_version()

public Bio::EnsEMBL::Hive::Meadow::get_meadow_major_version

(

)

Undocumented method

Code:

get_meadow_version()

public Bio::EnsEMBL::Hive::Meadow::get_meadow_version

(

)

Undocumented method

Code:

get_report_entries_for_process_ids()

public Bio::EnsEMBL::Hive::Meadow::get_report_entries_for_process_ids

(

)

    Title   :  get_report_entries_for_process_ids
    Function:  A higher-level method that gets process_ids as input and
               returns a structure like parse_report_source_line.

Code:

get_report_entries_for_time_interval()

public Bio::EnsEMBL::Hive::Meadow::get_report_entries_for_time_interval

(

)

    Title   :  get_report_entries_for_time_interval
    Function:  A higher-level method that gets a time interval as input and
               returns a structure like parse_report_source_line.

Code:

job_array_common_name()

public Bio::EnsEMBL::Hive::Meadow::job_array_common_name

(

)

    Title   :  job_array_common_name
    Function:  More specific version of job_name_prefix() that returns
               the actual name that agents should have at a specific
               beekeeper loop.

Code:

job_name_prefix()

public Bio::EnsEMBL::Hive::Meadow::job_name_prefix

(

)

    Title   :  job_name_prefix
    Function:  Tells how the agents (workers) should be generally named. It
               is used to name new agents, and to find our own agents.

Code:

kill_worker()

public Bio::EnsEMBL::Hive::Meadow::kill_worker

(

)

    Title   :  kill_worker
    Function:  Kill a worker.

Code:

name()

public Bio::EnsEMBL::Hive::Meadow::name

(

)

    Title   :  name
    Function:  Returns the name of the Meadow (which excludes the Meadow type)

Code:

new()

public Bio::EnsEMBL::Hive::Meadow::new

(

)

    Title   :  new (constructor)
    Function:  Instantiates a new Meadow object

Code:

parse_report_source_line()

public Bio::EnsEMBL::Hive::Meadow::parse_report_source_line

(

)

    Title   :  parse_report_source_line
    Function:  Opens and parses a file / command-line to return the
               resource-usage of some workers. Should return a hashref
               where process_id is the key to a hashref composed of:
                 when_died
                 pending_sec
                 exception_status
                 cause_of_death
                 lifespan_sec
                 mem_megs
                 cpu_sec
                 exit_status
                 swap_megs

Code:

pipeline_name()

public Bio::EnsEMBL::Hive::Meadow::pipeline_name

(

)

    Title   :  pipeline_name
    Function:  Getter/setter for the name of the current pipeline.
               This method is used by other Meadow methods such as
               job_name_prefix().

Code:

run_on_host()

public Bio::EnsEMBL::Hive::Meadow::run_on_host

(

)

    Title   :  run_on_host
    Function:  Runs an arbitrary commands on the given host. The host is expected to belong to the meadow and be reachable

Code:

runWorker_path()

public Bio::EnsEMBL::Hive::Meadow::runWorker_path

(

)

    Title   :  runWorker_path
    Function:  Getter for the path to runWorker.pl
               This is now set in the JSON config file. When missing or set to null,
               defaults to $EHIVE_ROOT_DIR/scripts

Code:

signature()

public Bio::EnsEMBL::Hive::Meadow::signature

(

)

    Title   :  signature
    Function:  The "signature" of a Meadow is its unique identifier across
               the Valley.

Code:

status_of_all_our_workers()

public Bio::EnsEMBL::Hive::Meadow::status_of_all_our_workers

(

)

    Title   :  status_of_all_our_workers
    Function:  Returns an arrayref of arrayrefs [worker_pid, meadow_user, status, rc_name]
               listing the workers that this Meadow can see.
               Allowed statuses are "RUN", "PEND", "SSUSP", "UNKWN"

Code:

submit_workers_return_meadow_pids()

public Bio::EnsEMBL::Hive::Meadow::submit_workers_return_meadow_pids

(

)

    Title   :  submit_workers_return_meadow_pids
    Function:  Submit $required_worker_count workers with the command $worker_cmd and return the meadow-specific worker_pids

Code:

type()

public Bio::EnsEMBL::Hive::Meadow::type

(

)

    Title   :  type
    Function:  The "type" of a Meadow is basically its job management
               system. eHive comes with two Meadows: Platform LSF (type
               "LSF"), and a default fork()-based (type "LOCAL"). Other
               meadows can be implemented provided that they follow the
               right interface.

Code:

---

The documentation for this class was generated from the following file:

• modules/Bio/EnsEMBL/Hive/Meadow.pm