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

as needed. This is important when values are reset on SIGHUP.

(gnats:381)

to relinquish control. This is necessary since a temporary network problem
can result in the BackupController becoming the primary server even while
the server on the ControllerMachine continues execution. While this event
is impossible to prevent, the new code restores proper operation when
communications are restored.
(gnats:387)

state recovery.

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.

  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).

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.

slurmctld to exit.

  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

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

This was a problem briefly observed in RedHat 9.

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.

then don't treat as fatal error.

modifications were relatively minor - mostly changes in function
names or arguments.

  (it was inappropriately going to DRAINING state).

  see memory leaks.

  (it grows anyway).
Fix state read logic to better handle error conditions.

Fix update node RPC to handle reason field change without state change.
  State was being handled as type int instead of uint16_t so NO_VAL check
  was not working properly.

"Scontrol abort" works. It was leaving a hung pthread due to a recent
   change.
Fix a couple of potential memory leaks
"switch_type" has been added to config data structure, un/pack, etc,
   but not yet reported to the user or documented yet.
The plugins now use function calls to get a their type and plugin directory
   from a common data structure rather than individually reading and
   parsing the configuration file.