Arguments

managed-job-globusrun [options] [<factory>] <job description>
                managed-job-globusrun -p -file <job description filename>
                managed-job-globusrun (-state | -release | -kill) <job handle>
                managed-job-globusrun -help | -usage | -version

with

<job description>     = -file <job description filename> | <command line>
                <factory> = -factory <contact> [-type <type>]
                <contact> = [<protocol>://]<host>[:[port]][/<service>]
                [options] = [-q] [-n]
                [-b] [-duration] [-terminate-at]
                [-auth <auth>] [-xmlsec <sec>] [-personal]
                [-submission-id <ID>]
            

Options

'executable (argument)*'

-terminate-at '08/15/2005 11:30'

Substitution variables

In GT 3.9.2, job description substitution variables had been removed from GRAM. Starting with GT 3.9.5, substitution variables are available again, while preserving the simplicity of the job description XML schema (relative to the GT3.2 job description schema). Substitution variables can be used in any path-like string or URL specified in the job description. They are special strings that are replaced by the GRAM services with actual values that the client-side does not a priori know. An example of substitution variable is ${GLOBUS_USER_HOME}, which represents the path to the HOME directory on the file system visible by the GRAM services of the user on behalf of whom the job is executed.

Details are in job description doc

Submission ID

A submission ID may be used in the GRAM protocol for robust reliability in the face of message faults or other transient errors in order to ensure that at most one instance of a job is executed, i.e. to prevent accidental duplication of jobs under rare circumstances with client retry on failure. The managed-job-globusrun tool always uses this feature, requiring either a submission ID to be passed in as input or a new unique ID to be created by the tool itself. If a new ID is created, it should be captured by the user who wishes to exploit this reliability interface. The ID in use, whether created or passed as input, will be written to the first line of standard output unless the quiet mode is in effect.

If a user is unsure whether a job was submitted successfully, he should resubmit using the same ID as was used for the previous attempt.

Job hold and release

It is possible to specify in a job description that the job be put on hold when it reaches a chosen state (see GRAM Approach documentation for more information about the executable job state machine, and the job description XML schema documentation for information about how to specify a held state This is useful for instance when a GRAM client wishes to directly access output files written by the job (as opposed to waiting for the stage-out step to transfer files from the job host). The client would request that the file cleanup process be held until released, giving the client an opportunity to fetch all remaining/buffered data after the job completes but before the output files are deleted.

Note that the hold feature of the GRAM service interface is not exploited by the current Java version of the client tool, but will be in the C client in order to implement client-side streaming of remote stdout/err.

The current client tool does however

MultiJobs

The new job description XML schema allows for specification of a MultiJob i.e. a job that is itself composed of several executable jobs (those jobs cannot be multijobs, so the structure is not recursive). This is useful in order to bundle a group of jobs together and submit them as a whole to a remote GRAM installation.

Note that there is no specification of relationships between the executable jobs, which we will refer to as "subjobs". The subjobs are submitted to job factory services in their order of appearance in the multijob description.

Job and process rendezvous

This version of GRAM offers a mechanism to perform synchronization between job processes in a multiprocess job and between subjobs in a multijob. The job application can in fact register binary information, for instance process information or subjob information, and get notified when all the other processes or subjobs have registered their own information. This is for instance useful for parallel jobs which need to rendezvous at a "barrier" before proceeding with computations, in the case when no native application API is available to help do the rendezvous.

No more file staging using GASS

The GASS server is not being used anymore by GRAM, so the options -server and -write have been removed. Instead, file staging is done in a reliable fashion via RFT and GridFTP servers. file staging in GT 4.0 GRAM

No standard output redirection yet

Unlike the GT3.2 managed-job-globusrun used with the option -output, this version of the tool does not offer any streamed redirection of the standard streams. This is because the GASS server is not used anymore by GRAM. Instead, a future version of the tool will allow for streaming of any server-side file (including the standard streams of the job execution) using GridFTP "tailing" of remote files.

No listing of submitted jobs yet

The -list option, which made the 3.2 tool print the identifiers of the jobs submitted by the user on the standard output, is not available in this version of the tool.

Tool-triggered automatic job resource destruction

Execution errors and user interrupt events are handled by automatically destroying the requested job service(s), unless the -batch option is on the command-line. The -batch option prevents the tool from listening to job state changes and from waiting for the job to finish. If -batch is selected, the command will return as soon as the remote job has been submitted.

The behavior of the tool with respect to job service destruction will vary in response to several kinds of events:

If the Java virtual machine of the tool aborts, that is, stops running without shutting down cleanly, for instance because it received a SIGKILL signal on Unix, then no guarantee can be made about whether or not the job service(s) will be destroyed.

Note: the shutdown behavior explained above cannot be guaranteed if the JVM option -Xrs is entered. The recommended way to disable service destruction is to specify the -batch option on the command-line.

Credential delegation

Submitting a job in interactive mode

A very simple command-line can be used to submit a job. For instance, the following command-line submits a job to the GRAM services hosted on the same machine (assuming a Globus container is running of course):

% bin/managed-job-globusrun "/bin/echo Testing 1...2...3"

The output should look like:

                Submission ID: uuid:661AA7F0-2573-11D9-99B2-D4755757F903
                WAITING FOR JOB TO FINISH
                ========== State Notification ==========
                Job State: Active
                ========================================
                ========== State Notification ==========
                Job State: CleanUp
                ========================================
                ========== State Notification ==========
                Job State: Done
                ========================================
                Exit Code: 0
                DESTROYING SERVICE
                SERVICE DESTROYED
            

Note: the job state notifications are printed in the order of arrival, but they may arrive at the client-side in any order.

In this example the job description specifies the standard output stream path of the job to be: ${GLOBUS_USER_HOME}/stdout. The GRAM services replace the substitution variable ${GLOBUS_USER_HOME} with the path to the Home directory of the submitting user as seen by the machine were the invoked GRAM services are hosted. You can thus verify the output of the job with the following command:

% cat ~/stdout

which will display the string:

12 abc 34 pdscaex_instr_GrADS_grads23_28919.cfg pgwynnel was here

Submitting a job in batch mode, checking its status and destroying the resource

To submit a job without having the client wait for job completion, specify the option -batch (or -b) on the command-line:

                % bin/managed-job-globusrun -batch "/bin/echo Testing 1...2...3"
                Warning: Will not wait for job completion, and will not destroy job service.
                Submission ID: uuid:9C715240-26C7-11D9-850A-ABE2020F9ED6
                CREATED MANAGED JOB SERVICE WITH HANDLE:
                http://127.0.0.1:8080/wsrf/services/ManagedExecutableJobService?9C715240-26C7-11D9-850A-ABE2020F9ED6
            

To check the status of the job, use the -state option:

                % bin/managed-job-globusrun 
                -state 'http://127.0.0.1:8080/wsrf/services/ManagedExecutableJobService?9C715240-26C7-11D9-850A-ABE2020F9ED6'
                Job State: Done
            

To destroy the job resource created on the server side, use the -kill option:

                % bin/managed-job-globusrun 
                -kill 'http://127.0.0.1:8080/wsrf/services/ManagedExecutableJobService?9C715240-26C7-11D9-850A-ABE2020F9ED6'
                DESTROYING SERVICE
                SERVICE DESTROYED
            

Finding which schedulers are interfaced by the WS GRAM installation

Unfortunately there is no option yet to print the list of local resource managers supported by a given WS-GRAM service installation. But there is a way to check, whether WS-GRAM supports a certain local resource manager or not. The following command gives an example of how a client could check if Condor is available at the remote site:

wsrf-query \
    -s https://<hostname>:<port>/wsrf/services/ManagedJobFactoryService 
    -key {http://www.globus.org/namespaces/2004/10/gram/job}ResourceID Condor \
    "//*[local-name()='version']"
              

Replace host and port settings with the values you need. If Condor is available on the server-side, the output should look something like the following:

	<ns1:version xmlns:ns1="http://mds.globus.org/metadata/2005/02">4.0.3</ns1:version>
   

In this example the output indicates, that a GT is listening on the server-side, that Condor is available and that the GT version is 4.0.3. If no GT at all is running at the specified host and/or port or if the specified local resource manager is not available on the server-side, the output will be an error message.

On the server-side the GRAM nameof local resource managers for which GRAM support has been installed can be obtained by looking at the GRAM configuration on the GRAM, as explained here The GRAM name of the local resource manager can be used with the -type option to specify which factory resource to use when submitting a job. For instance:

                    % bin/managed-job-globusrun -type Fork /bin/true
                

will submit a /bin/true job to the Fork local resource manager (i.e. the command-line /bin/true will simply be executed as a newly spawn process)

                    % bin/managed-job-globusrun -type LSF /bin/true
                

will submit a /bin/true job to the LSF scheduler (if installed).

                    % bin/managed-job-globusrun -type Multi -file simple_multi_job.xml
                

where simple_multi_job.xml contains the description of a multijob will submit a multi job to the Multi ManagedJobFactory resource.

Specifying file staging in the job description

In order to do file staging one must add specific elements to the job description. The file transfer directives follow the RFT syntax RFT syntax ,which enables third-party transfers. Each file transfer must therefore specify a source URL and a destination URL. URLs are specified as GridFTP URLs (for remote files) or as file URLs (for local files).

For instance, in the case of staging a file in, the source URL would be a GridFTP URL (for instance gsiftp://job.submitting.host:2888/tmp/mySourceFile) resolving to a source document accessible on the file system of the job submission machine (for instance /tmp/mySourceFile). At run-time the Reliable File Transfer service used by the GRAM service on the remote machine would fetch the remote file using the GridFTP protocol and write it reliably to the specified local file (for instance file:///${GLOBUS_USER_HOME}/my_transfered_file, which resolves to ~/my_transfered_file). Here is how the stage-in directive would look like:

                    <fileStageIn>
                    <transfer>
                    <sourceUrl>gsiftp://job.submitting.host:2888/tmp/mySourceFile</sourceUrl>
                    <destinationUrl>file:///${GLOBUS_USER_HOME}/my_transfered_file</destinationUrl>
                    </transfer>
                    </fileStageIn>
                

Note: additional RFT-defined quality of service requirements can be specified for each transfer. See the RFT documentation for more information.

Here is an example job description with file stage-in and stage-out:

                    <job>
                    <executable>my_echo</executable>
                    <directory>${GLOBUS_USER_HOME}</directory>
                    <argument>Hello</argument>
                    <argument>World!</argument>
                    <stdout>${GLOBUS_USER_HOME}/stdout</stdout>
                    <stderr>${GLOBUS_USER_HOME}/stderr</stderr>
                    <fileStageIn>
                    <transfer>
                    <sourceUrl>gsiftp://job.submitting.host:2888/bin/echo</sourceUrl>
                    <destinationUrl>file:///${GLOBUS_USER_HOME}/my_echo</destinationUrl>
                    </transfer>
                    </fileStageIn>
                    <fileStageOut>
                    <transfer>
                    <sourceUrl>file://${GLOBUS_USER_HOME}/stdout</sourceUrl>
                    <destinationUrl>gsiftp://job.submitting.host:2888/tmp/stdout</destinationUrl>
                    </transfer>
                    </fileStageOut>
                    <fileCleanUp>
                    <deletion>
                    <file>file://${GLOBUS_USER_HOME}/my_echo</file>
                    </deletion>
                    </fileCleanUp>
                    </job>
                

The submission of this job to the GRAM services causes the following sequence of actions:

Specifying and submitting a MultiJob

Within the multijob description, each subjob description must come along with an endpoint for the factory to submit the subjob to. This enables the at-once submission of several jobs to different hosts. The factory to which the multijob is submitted acts as an intermediary tier between the client and the eventual executable job factories. See the job description schema documentation for more information about multijob specification.

A multijob must be submitted to a Multi job factory resource:

                    % bin/managed-job-globusrun -type Multi -file myMultiJob.xml
                

A multijob resource is created by the factory and exposes a set of WSRF resource properties different than the resource properties of an executable job. The state machine of a multijob is also different since the multijob represents the overall execution of all the executable jobs it is composed of.

Job Execution Errors

fault types

Common issues

                    Error: error submitting job request: ; nested exception is:
                    javax.xml.rpc.soap.SOAPFaultException: Expired credentials 
                    (O=Grid,OU=GlobusTest,OU=simpleCA.foo.bar.com,OU=bar.com,CN=John Doe,CN=1255793213).
                

                    % bin/grid-proxy-init
                    Your identity: /O=Grid/OU=GlobusTest/OU=simpleCA-foo.bar.com/OU=bar.com/CN=John Doe
                    Enter GRID pass phrase for this identity:
                    Creating proxy ................................. Done
                    Your proxy is valid until: Tue Oct 26 01:33:42 2004
                

                    Error: error submitting job request: ; nested exception is:
                    java.net.SocketTimeoutException: Read timed out
                

                    % bin/managed-job-globusrun -timeout 240000 -f myJob.xml
                

                    Unable to create RFT Resource; nested exception is:
                    org.apache.commons.dbcp.DbcpException: Connection refused. Check that the hostname 
                    and port are correct and that the postmaster is accepting TCP/IP connections.
                

                    Error: error submitting job request: 
                    org.globus.wsrf.impl.security.authorization.exceptions.AuthorizationException: 
                    (pdp08) "/O=Grid/OU=GlobusTest/OU=simpleCA-foo.bar.com/OU=foo.bar.com/CN=John Doe" 
                    is not authorized to use operation:
                    {http://properties.impl.wsrf.globus.org}getMultipleResourceProperties on this service
                

                    [Thread-3] WARN  factory.ManagedJobFactoryResource [getRestartTimestamp:187] 
                    java.io.FileNotFoundException: /software/globus/gt4/rc4.0.0/var/globus-jsm-fork.stamp (No such file or directory)
                    [Thread-3] WARN  factory.ManagedJobFactoryResource [getRestartTimestamp:187] 
                    java.io.FileNotFoundException: /software/globus/gt4/rc4.0.0/var/globus-jsm-multi.stamp (No such file or directory)
                    [Thread-2] WARN  utils.XmlPersistenceHelper [load:185] [CORE] File 
                    /nfs/v5/alain/.globus/persisted/128.9.72.67/ManagedExecutableJobResourceStateType/897BC6E0-26CA-11D9-8D59-FF280F77E689.xml
                    for resource {http://www.globus.org/namespaces/2004/10/gram/job}ResourceID=897BC6E0-26CA-11D9-8D59-FF280F77E689 
                    was not found
                

Known problems

                    % rm var/globus-jsm-*.stamp
                

                    [Thread-7] ERROR jobmanager.JobManager [unsubscribeForNotifications:1762] unable to stop monitoring job for state changes
                    org.globus.exec.monitoring.NotRegisteredException
                    at org.globus.exec.monitoring.JobStateMonitor.unregisterJobID(JobStateMonitor.java:375)
                    at org.globus.exec.service.job.jobmanager.JobManager.unsubscribeForNotifications(JobManager.java:1758)
                    at org.globus.exec.service.job.jobmanager.JobManager.processState(JobManager.java:1274)
                    at org.globus.exec.service.job.jobmanager.RunQueue.run(RunQueue.java:75)