Job Files 
Jobs generate output, usually at least one output file and when errors occurred there is also a log file. These files can be plain text files or of another format , like PDF, depending on the job. These files are stored on the process server, SAP and System job output gets stored on the Central Scheduler; Remote agents store output of jobs locally. Sufficient free space must be available on these file systems,
The default output file of jobs is named stdout, which refers to the concept of the standard output stream of computer systems. The log file is named stderr which refers to the concept of standard error stream of computer systems. Additional output files have the name of the output, if for example in a KSH or CMD job definition you direct output to a file, the filename will be the name of the output.
In the following example, the job file will have the name of the output, done.txt.
echo done > done.txt
The location of the output and log files is configured with the DataRootDirectroy process server parameter, which can be set for all process servers. The DataRootDirectory also contains process server logfiles, as well as other data used by file event processes.
Note
The process server parameter can have a trailing slash / or \, however, this is not required.
Under the DataRootDirectroy a lis and log directory will be created, the first for the output, which by default will have a txt extension, and the latter for log files, which by default will have a log extension. The extension of the job files depends on the output and log format.
SAP and System Process Servers
SAP and System process servers do not have the parameter in their parameters list and use the default value, which is java.io.tmpdir, and depends on the configuration of your J2ee Application Server, where it can be changed centrally.
In the Windows Standalone Edition, the default java.io.tmpdir, is set to ${InstallDir}\j2ee\cluster\server1\temp.
In SAP NetWeaver, this value is set to the system temp directory by default, which is %windir%\temp on Windows by default and /tmp on most Unix systems by default.
The DataRootDirectory of the System process server can be set with the job definition System_SetSystemDataRootDirectory.
The DataRootDirectory of SAP process servers just needs to be set on the process server. Previous versions required the process server to have a System job definition type, this is no longer the case.
Platform Agents
The default value of DataRootDirectroy is set to <InstallDir>/var by default, this value should be changed to a directory on a large file system.
<InstalLDir> is the installation directory, on Windows the default is %ProgramFiles%\Redwood\Server on Unix systems there is no default, it will default to the current working directory.
Note
Previous releases had the DataRootDirectory set to <InstallDir>/<version>/var; Redwood highly recommends to set the DataRootDirectory process server parameter before upgrading if you had not set it previously. If you choose not to, file events can loose their state after the upgrade.