Class IMonitoringService

Initialize the monitoring service.

If the monitoring service is created in the Ganga client then job_info is the original job object, and config_info is the original config object returned by getConfig().

If the monitoring service is created on the worker node then job_info is the return value of getJobInfo(), and config_info is the return value of getConfig().getEffectiveOptions(), i.e. a dictionary.

In order to support existing monitoring classes, which do not use config_info, if getConfig() returns None, the constructor is called with only the job_info argument.

Completion of a job. Called by: ganga client.

Failure of a job. Called by: ganga client.

Return the config object for this class.

By default this method returns None. If it is overridden to return a Config object then the configuration is made available consistently in the Ganga client and on the worker node via the instance variable config_info. See __init__().

For example, in your IMonitoringService implementation you could add:

   def getConfig():
       from Ganga.Utility import Config
       return Config.getConfig("MyMS")
   getConfig = staticmethod(getConfig)

N.B. this is a static method.

Return a static info object which static information about the job at submission time. Called by: ganga client.

The info object is passed to the contructor. Info object may only contain the standard python types (such as lists, dictionaries, int, strings).

Get the list of module dependencies of this monitoring module.
Called by: ganga client.

Returns a list of modules which are imported by this module and
therefore must be shipped automatically to the worker node. The list
should include the module where this class is defined plus all modules
which represent the parent packages. The module containing the
IMonitoringService class is added automatically by the call to the
base class sandBoxModule() method. An example for a class defined in
the module Ganga/Lib/MonitoringServices/DummyMS/DummyMS.py which does
not have any further dependencies:

import Ganga.Lib.MonitoringServices.DummyMS
return IMonitoringService.getSandboxModules(self) + [
         Ganga,
         Ganga.Lib,
         Ganga.Lib.MonitoringServices,
         Ganga.Lib.MonitoringServices.DummyMS,
         Ganga.Lib.MonitoringServices.DummyMS.DummyMS
        ]
Note, that it should be possible to import all parent modules without side effects (so without importing automatically their other children).

Return a line of python source code which creates the instance of the monitoring service object to be used in the job wrapper script. This method should not be overriden.

Killing of a job. Called by: ganga client.

Preparation of a job. Called by: ganga client.

Application execution is in progress (called periodically, several times a second). Called by: job wrapper.

Rollback of a job to new state (caused by error during submission). Called by: ganga client.

Application is about to start on the worker node. Called by: job wrapper.

Application execution finished. Called by: job wrapper.

Submission of a job. Called by: ganga client.

Just before the submission of a job. Called by: ganga client.