Contents

1  Introduction to data analysis
    1.1  Getting Started
    1.2  Getting data
    1.3  Replaying data

List of Tables

1.  Introduction to data analysis

1.1  Getting Started

You should get a copy of the ESPACE manual (Link to espace information).  Need to have group permissions for a-e93027. To check if you have group permissions type groups at the command prompt. If you don't have group permission then just send email to Bob Michaels asking him to put you in the group. If using csh then edit your .cshrc and add the following lines

setenv HALLA_DIR  /work/halla/software

setenv HALLA_BIN $HALLA_DIR/bin

setenv LOC_HALLA_BIN $HALLA_BIN/$OSNAME

setenv PATH $LOC_HALLA_BIN/:$PATH

You can type source .cshrc to set the above variables without a logout and login. In your home directory execute the command cp -R  /work/halla/e93027/cosmic espace. This will create an analysis directory. Create the directory /work/halla/e93027/name/hbook where name is whatever you want. Then type the command ln -s /work/halla/e93027/name/hbook espace. This creates a symbolic link between your espace/hbook directory and where the espace kumac will put the hbook files. Just makes it easier to find things in paw++. The work disk is not backed up so it is recommended to work on the home disk which is backed up and store hbook files on the work disk. If you want to save the hbook files then copy them to /mss/home/username .

1.2  Getting data

Data is stored on the mass storage system (mss, link to information of mss). The way to get data off the MSS is to do the following command

setup mss

request_file -f   /mss/halla/ exp/raw/ exp_####.da t

exp is the experiment number, for example e89003, e89033, e93050, e93027 etc. The file will be taken from the MSS and put in the directory /cache/halla/exp/raw . An email will be sent to you when this process is complete. One needs to put the data on the cache disk to replay the data interactively either on the jsub machines (not recommended, very slow) or on the ifarm machines. When replaying data in batch mode on the farm machines one can get the data file directly from the mss and not have to put it on the cache disk. There are also the commands jget and jput which can be used to move data to and from the work or home disk to the mss and also get files from the cache disk. ( More information of batch commands)

1.3  Replaying data

Data can be replayed interactively on the jlab (for very short test s) or ifarm (for longer tests) machines. When wanting to replay interactively move to /home/username/espace and type espace_fpp (this version of espace is usually only slightly different than the official version of espace with the latest changes having to do with fpp parts of espace). You should get an espace> prompt. At the prompt type exec espace_kumac/fpp_offline.kumac or which ever espace kumac you want to run. After running espace will create an hbook file in your hbook directory.

Replaying data on the batch farms machines (farms1, farms2 etc) can be done using the jsub command. (More details on batch farm at JLAB). The jsub command needs a script file and also a command file to run espace on the batch computer. A command file has been written called replay_fpp which creates the jsub script file, the espace command file and executes the jsub command. To check the status of the batch job use the command jobstat .

File translated from TEX by TTH, version 0.9.