of slurmd process.

    calls, improves performance and eliminates realloc failure on
    RH EL3 under extremely heavy workload apparently due to memory
    fragmentation

Requires setting counts in updated batch job launch request message.

    corresponding file descriptor for readability. Otherwise, srun
    and slurmd will spin when stdin file descriptors are full because
    pending data will always go unread. (gnats:434)

specifying an infinite time limit.

are specific to a particular plugin type.

commands modified. Info filtering added to slurmctld.

loadable as shared aliases. Give slurm functions a prefix of "slurm_" to
insure that no user functions with the same name is used instead.
[For some reason a this same commit a minute ago did not seem to take effect.]

right away and treat failure as a fatal error.

write to.

    slurmctld startup
Create StateSaveLocation directory if changes via slurmctld reconfig

Properly support sinfo --state option value including "*" suffix.

  (which exhaust available memory).
Restructure slurmctld agent logic for higher throughput.

Fixes problem running very large number of simultaneous jobs when
their inactivity limit was reached due to a backlog of message.

data structure. This eliminates risks associated with re-reading slurm.conf.

them when partition limits change.

the plugins themselves. Add new functions to each plugin to return the
error number and given an error number, return a description.

  parameter.