and when reading in state from DB2 we find a block that can't be created.
You can now do a clean start to rid the bad block.

the slurmctld there were software errors on some nodes.

without it still existing there.  This is extremely rare.

a pending job on it we don't kill the job.

while it was free cnodes would go into software error and kill the job.

Fix for bug 224

WIthout this patch, when the slurmd cold starts or slurmstepd
terminates abnormally, the job script file can be left around.
bug 243

Previously such a job submitted to a DOWN partition would be queued.
bug 187

Fix spank_option_getopt in local context

In local context (srun, sbatch, salloc), spank_option_getopt() would
always return that an option was found due to a missing check for
spopt->found in spank_option_getopt.

This patch fixes the issue.

slurm.schedmd.com

a reservation when it has the "Ignore_Jobs" flag set.  Since jobs could run
outside of the reservation in it's nodes without this you could have
double time.

This is not applicable with launch/aprun

On Redhat 6 based distros the lua library name is liblua-5.1.so.
Installing the lua-devel package will create the liblua.so symlink, but
if that isn't installed then the lua job submit plugin will fail to
load.
I'm attaching a patch that adds liblua-5.1.so to the search path.

These changes are required so that select/cray can load select/linear,
  which is a bit more complex than the other select plugin structures.
Export plugin_context_create and plugin_context_destroy symbols from
  libslurm.so.
Correct typo in exported hostlist_sort symbol name
Define some functions in select/cray to avoid undefined symbols if
  the plugin is loaded via libslurm rather than from a slurm command
  (which has all of the required symbols)