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

  parameter.

  Attempted "scontrol reconfig" or SIGHUP requests result in error and
  ignoring change to SwitchType. Document how to change a cluster's SwitchType.

  due to multiple background theads that are difficult to clean-up.

daemon restart).

reconfiguration (rather than restart).

the directory with access rights based upon its umask. A umask value of
007 when creating the directory results in batch jobs always failing.

responding. Wait for tasks to complete or administrator to set DOWN.

If any errors occur unpacking a job credential, don't recover that job
step or any subsequent jobs.

Slurmctld must be restarted to re-load any plugin.

in addition to the formerly supported alphabetical order. Reporting
partitions in configuration order is the new default behavior.

Add LDFLAGS required on AIX to use dlopen().

Leaving the work to someone else for now...

slurmctld to exit.

"API" to META file.

Export version info via new API

  clusters with FastSchedule configured off
* Only return DOWN nodes to service if the reason for them being in
  that state is non-responsiveness and ReturnToService configured on
* Some general code clean-up

displayed (e.g. cpus, disk, memory, or weight). Formerly it separated
out nodes with differing values for those fields even if not displayed.

fast_schedule == 0 (i.e. the node's value is used directly rather
than the partition's configuration value).

Local code added that is built and used only as needed.

lacking an elan switch.

This was a problem briefly observed in RedHat 9.

alloca() has been replace by malloc() and free(),
Use of __progname eliminated,
GNU compliant malloc() wrapper added,
RLIMIT_* definition use controlled by #ifdef's,
Changed some data types, and conditionally added some #includes.

was being ignored by default when manually initiated, which prevents it
from terminating gracefully.

when a node becomes DOWN for not responding. This is because if there
are a large number of non-responsive nodes, the ping agent can take a
long time to complete (one second per non-responsive node or 10 second
timeout per node with 10 parallel tasks). This should more properly mark
nodes as DOWN.

than two with different naming convensions and values.

 o Do not restore saved credential state when a clean start is requested.

then don't treat as fatal error.

 o change make_time_str to slurm_make_time_str in libslurm for
   consistency and safety.
 o update references to make_time_str to slurm_make_time_str