DAG job

From MediaWiki

Revision as of 15:33, 27 May 2012 by Ngrkic (Talk | contribs)
Jump to: navigation, search

Submitting DAG job how-to


This is DAG job submiting example. For demonstration use attached file. In this case job is split in 4 independent jobs. First A job will be executed, then B and C jobs, and finally job D

Dag2.png

DAG.tgz

1.Extract file with:

[ngrkic@ui ~]$ tar xvzf DAG.tgz

2.Enter directory single with:

[ngrkic@ui ~]$ cd DAG

3.List directory:

[ngrkic@ui DAG]$ ll

-rw-r--r--  1 ngrkic ngrkic 1469 Oct  1 15:54 jobDAG.jdl
-rwxr-xr-x  1 ngrkic ngrkic   92 Oct  1 15:54 job.sh

This is content of file jobDAG.jdl

[

Type = "dag"; InputSandbox = {"job.sh", "job2.sh"};

Nodes = [

nodeA = [
          Description = [
           JobType = "Normal";
           Executable = "job.sh";
           Arguments = "A";
           StdOutput = "std.out";
           StdError = "std.err";
           InputSandbox = {root.InputSandbox[0]};
           OutputSandbox = {"std.out","std.err"};
          ];
        ];
nodeB = [
          Description = [
           JobType = "Normal";
           Executable = "job2.sh";
           Arguments = "B";
           StdOutput = "std.out";
           StdError = "std.err";
           InputSandbox = {root.InputSandbox[1]};
           OutputSandbox = {"std.out","std.err"};
          ];
        ];
nodeC = [
          Description = [
           JobType = "Normal";
           Executable = "job3.sh";
           Arguments = "C";
           StdOutput = "std.out";
           StdError = "std.err";
           InputSandbox = {"job3.sh"};
           OutputSandbox = {"std.out","std.err"};
          ];
        ];
nodeD = [
          Description = [
           JobType = "Normal";
           Executable = "job.sh";
           Arguments = "D";
           StdOutput = "std.out";
           StdError = "std.err";
           InputSandbox = {root.InputSandbox[0]};
           OutputSandbox = {"std.out","std.err"};
          ];
        ];

]; Dependencies = {

               {nodeA,nodeB},{nodeA,nodeC},{{nodeB,nodeC},nodeD}

}; ]

This is content of file job.sh

#!/bin/bash 

echo "Job $1 - `date` - BEGIN"
hostname
sleep 100
echo "Job $1 - `date` - END"


4.Creating VOMS proxy:

[ngrkic@ui DAG]$ voms-proxy-init -voms aegis
Cannot find file or dir: /home/ngrkic/.glite/vomses
Enter GRID pass phrase:
Your identity: /C=RS/O=AEGIS/OU=Institute of Physics Belgrade/CN=Nikola Grkic
Creating temporary proxy ......................... Done
Contacting  voms.ipb.ac.rs:15001 [/C=RS/O=AEGIS/OU=Institute of Physics Belgrade/CN=host/voms.ipb.ac.rs] "aegis" Done
Creating proxy ..................................................... Done
Your proxy is valid until Fri Oct  2 01:11:00 2009


5.Submiting DAG job:

[ngrkic@ui DAG]$ glite-wms-job-submit -a jobDAG.jdl 
Connecting to the service https://wms-aegis.ipb.ac.rs:7443/glite_wms_wmproxy_server


====================== glite-wms-job-submit Success ======================

The job has been successfully submitted to the WMProxy
Your job identifier is:

https://wms-aegis.ipb.ac.rs:9000/O3Bxi2I9DoE2Lltmcoe0eQ

==========================================================================


Copy the job ID.Job is running now, and it should finish in few moments...


6.Requesting Job status:

[ngrkic@ui DAG]$ glite-wms-job-status https://wms-aegis.ipb.ac.rs:9000/O3Bxi2I9DoE2Lltmcoe0eQ


*************************************************************
BOOKKEEPING INFORMATION:

Status info for the Job : https://wms-aegis.ipb.ac.rs:9000/O3Bxi2I9DoE2Lltmcoe0eQ
Current Status:     Done (Success)
Exit code:          0
Status Reason:      Job terminated successfully
Destination:        dagman
Submitted:          Thu Oct  1 16:44:32 2009 CEST
*************************************************************

- Nodes information for: 
    Status info for the Job : https://wms-aegis.ipb.ac.rs:9000/4xL7Bu5cI6eaixg9y0oezg
    Current Status:     Done (Success)
    Logged Reason(s):
        - 
        - Job terminated successfully
    Exit code:          0
    Status Reason:      Job terminated successfully
    Destination:        grid-ce.etf.bg.ac.rs:2119/jobmanager-pbs-aegis
    Submitted:          Thu Oct  1 16:44:32 2009 CEST
*************************************************************
   
    Status info for the Job : https://wms-aegis.ipb.ac.rs:9000/LCE9ULBND3jYd2JUJPJ9wQ
    Current Status:     Done (Success)
    Logged Reason(s):
        - 
        - Job terminated successfully
    Exit code:          0
    Status Reason:      Job terminated successfully
    Destination:        grid01.rcub.bg.ac.rs:2119/jobmanager-pbs-aegis
    Submitted:          Thu Oct  1 16:44:32 2009 CEST
*************************************************************
    
    Status info for the Job : https://wms-aegis.ipb.ac.rs:9000/quIESEluISRrAXdcC7ELKQ
    Current Status:     Done (Success)
    Exit code:          0
    Status Reason:      Job terminated successfully
    Destination:        grid01.elfak.ni.ac.rs:2119/jobmanager-pbs-aegis
    Submitted:          Thu Oct  1 16:44:32 2009 CEST
*************************************************************
    
    Status info for the Job : https://wms-aegis.ipb.ac.rs:9000/yNu3yqhSwf4sZ25bWuzKDg
    Current Status:     Done (Success)
    Exit code:          0
    Status Reason:      Job terminated successfully
    Destination:        grid01.elfak.ni.ac.rs:2119/jobmanager-pbs-aegis
    Submitted:          Thu Oct  1 16:44:32 2009 CEST
*************************************************************

7.Requesting Job output:

[ngrkic@ui DAG]$ glite-wms-job-output --dir /home/ngrkic/test https://wms-aegis.ipb.ac.rs:9000/O3Bxi2I9DoE2Lltmcoe0eQ
Connecting to the service https://wms-aegis.ipb.ac.rs:7443/glite_wms_wmproxy_server

================================================================================

			JOB GET OUTPUT OUTCOME

Output sandbox files for the DAG/Collection :
https://wms-aegis.ipb.ac.rs:9000/O3Bxi2I9DoE2Lltmcoe0eQ
have been successfully retrieved and stored in the directory:
/home/ngrkic/test

================================================================================


8.Go to test directory and see output with commands:

[ngrkic@ui ~]$ cd 
[ngrkic@ui ~]$ cd test
[ngrkic@ui test]$ ll

drwxr-xr-x  2 ngrkic ngrkic 4096 Oct  1 17:31 nodeA
drwxr-xr-x  2 ngrkic ngrkic 4096 Oct  1 17:31 nodeB
drwxr-xr-x  2 ngrkic ngrkic 4096 Oct  1 17:31 nodeC
drwxr-xr-x  2 ngrkic ngrkic 4096 Oct  1 17:31 nodeD
[ngrkic@ui test]$ cd nodeA
[ngrkic@ui nodeA]$ ll
-rw-rw-r--  1 ngrkic ngrkic   0 Oct 12 17:31 std.err
-rw-rw-r--  1 ngrkic ngrkic 112 Oct 12 17:31 std.out

[ngrkic@ui nodeA]$ cat std.out 
Job A - Thu Oct 1 17:14:14 CEST 2009 - BEGIN
grid10.elfak.ni.ac.rs
Job A - Thu Oct 1 17:14:54 CEST 2009 - END

For each node there is similar output.

Personal tools