2.1.2 Getting the input files from environment variablesĪnother way to retrieve the file list is to look for the $INPUTFILECOUNT You have an output file for each input file, and all your input files haveĭifferent names, you can have the output file depend on the particular input.Īll the input files are assigned to only one process. The job ID, which you can retrieve from the environment variable $JOBID. You can, forĮxample, have the name of the output files depend on the filelist name, or from The standard out all the input files that were assigned to that particularīe careful that the output file is different for every process it might beĭispatched, otherwise two processes that were dispatched to different machinesĪnd are running in the same moment will interfere with each other. For example, submitting cat $FILELIST will display in You can use this variable on the command line to retrieve the name as a The environment variable $FILELIST will be set to the name of the file list. The scheduler will generate a file list like this for each process: /star/data21/reco/productionCentral/FullField/P02gc/2001/312/st_physics_2312011_raw_ 2.1.1 Getting the input files from the filelist For your convenience, there are two ways this is done.
Program will have to receive instruction from the scheduler which input files it What will you basically submit to the scheduler is a command line.
#Ischedule star manual
In this manual we show the steps the user needs to take to use the scheduler 2. XML file describing the job he wants to be done. User has to follow some simple rules when writing his program, and to write an In order for the scheduler to do that, the
The scheduler will divide the job into different processes that will beĭispatched to different machines. Input files, residing on NFS or on the various local disks of the farm machines. The STAR scheduler was developed to allow users to submit a job on several STAR scheduler - User manual STAR scheduler - User Manual 1.
0 kommentar(er)
