jlimit_startjob(3c)

NAME
     jlimit_startjob - creates a new job container

SYNOPSIS
     #include <sys/resource.h>

     jid_t    jlimit_startjob(char *username, uid_t uid, char *domainname);

DESCRIPTION
     The jlimit_startjob function performs all of the tasks necessary to
     creating a new job.

     The function takes the username, the user's ID (uid), and job limits
     domainname as arguments and returns the newly created job's job ID.

     A list of domains may also be passed to the jlimit_startjob function.
     This allows a batch scheduler to have a unique job limits domain.  The
     example below shows a domain list for the Miser batch processing system:

          jid = jlimit_startjob("bsmith", 7777, "miser:batch");


     The domain list is a colon-delimited list ("miser:batch").  The
     jlimit_startjob function starts with the first domain on the left; in
     this case, miser.  If the miser domain does not exist, then
     jlimit_startjob will try batch.

     To create a new job, the calling process must have CAP_SETUID and
     CAP_PROC_MGT capability.  See the capabilities(4) man page for more
     information on the capability mechanism that provides fine grained
     control over the privileges of a process.

     The jlimit_startjob function retrieves the job limit and process
     information from the ULDB and populates those limits into the job
     structure.

SEE ALSO
     showlimits(1), getjlimit(2), makenewjob(2).

ERRORS
     Under the following conditions, the function jlimit_startjob fails and
     set errno to:

     [ENOPKG]       The job limits software is not installed.

     [EINVAL]       The makenewjob(2) system call returns an invalid argument
                    to the jlimit_startjob function.

DIAGNOSTICS
     Upon successful completion, jlimit_startjob returns a job ID.  If an
     error occurs, a value of -1 is returned and errno is set to indicate the
     error.  If job limits software is not installed, jlimit_startjob returns
     a value of 0.