This improves the initial configuration code:
 a) Better handling of DownNodes lines
    The previous basil_geometry() would set the node Reason field on failure,
    irrespective of whether that node has been marked using a DownNode line.

 b) Check all cases of nodes being invisible to ALPS
    Up until now basil_geometry() had to be fixed each time a new source of
    discrepancy between ALPS and SDB state had been discovered (most recent
    case was NULL coordinates when taking out a blade). Depending on ALPS
    interface changes, there may be other possibilities. Instead of fixing the
    SLURM code for each new case, it is better to check whether SLURM and ALPS
    agree. The price is some tiny delay at SLURM initialisation time (since each
    node is first looked up in the ALPS inventory), but it pays well off as it
    eases system administration by pointing to the source of error.
    Any node that has suddenly disappeared from ALPS horizon will now show up in
    the logs, and also in marked down in sinfo.

 c) At initialisation time, give a summary as to how many ALPS nodes are online.

 d) Turn ALPS-node-invisibility error into warning message, since such nodes may
    already have been covered in a DownNodes statement.

By merging basil_get_initial_state() into basil_geometry(), the previously separate
knowledge about system state (database state, ALPS inventory) is combined, allowing
to more easily identify sources of failure.
Patch from Gerrit Renker, CSCS.

Latest Cray-specific modifiations

The enum is only needed and referenced in basil_geometry() and has otherwise
no special meaning since it directly depends on the selected output columns.
Patch from Gerrit Renker, CSCS.

This case was observed after taking a blade out of a CLE 2.x system. ALPS does not
list the removed nodes, but they still appear in the XTAdmin.processor table, with
NULL coordinates. Hence set node down if at least one coordinate is NULL.

Also add a check to compare how many out of the nodes in slurm.conf are visible to
ALPS (the absence of this test masked the bug), always list DOWN nodes at startup,
and clarify that not failing due to ALPS errors during the initial SLURM
configuration is not an option.

On the system which is missing a blade, the log information now is
 [2011-05-16T16:09:54] error: ALPS sees only 12/16 slurm.conf nodes
 [2011-05-16T16:09:54] Recovered state of 16 nodes
 [2011-05-16T16:09:54] Recovered state of 2 front_end nodes
 [2011-05-16T16:09:54] Recovered information about 0 jobs
 [2011-05-16T16:09:54] error: nid00028: unknown coordinates - hardware failure?
 [2011-05-16T16:09:54] error: nid00029: unknown coordinates - hardware failure?
 [2011-05-16T16:09:54] error: nid00030: unknown coordinates - hardware failure?
 [2011-05-16T16:09:54] error: nid00031: unknown coordinates - hardware failure?
Patch from Gerrit Renker, CSCS.

This fixes some errors in the documentation of how memory is allocated, and adds missing bits.
Patch from Gerrit Renker, CSCS.

BLUEGENE - Added block node cnt to be able to differentiate between a sub-block job and a regular full block job.

Conflicts:
	src/plugins/select/bluegene/bg_record_functions.c

BLUEGENE - if a block goes into an error state.  Fix issue where accounting wasn't updated correctly when the block was resumed.

Clearly document that only PreemptType=preempt/partition_prio can be
used with PreemptMode=suspend. Only partition data structures exist in
the module that suspends and resumes jobs.

The node state was formerly reported "UNKNOWN" on node state change
request errors.

This restores proper slurmd hostname resolution on Frontend systems
(BlueGene and Cray) which was broken by commit 635abaea
adding support for the slurmd -f option.

Remove local timer functions and make use of functions in src/common/timers.[ch]

Change diff_tv_str() to slurm_diff_tv_str() and
diff_tv() to slurm_diff_tv() so the symbols are exported
for use in the pmi library

Remove "inline" declarations on extern functions in src/common
so that the symbols are available to plugins.

When enforcing accounting, fix polling for unknown uids for users after the slurmctld started.  Previously one would have to issue a reconfigure to the slurmctld to have it look for new uids.

Reset priority to 0 when a job is requeued by wiki2

Conflicts:
	src/common/slurm_protocol_defs.h

Apply more patches to build SLURM on OSX. Patches from Tyler Strickland.

The job needs to be held by slurm until a STARTJOB command is
received from Moab.

Fix the setting of SLURM_TASKS_PER_NODE environment variable
in job step as needed by some MPI implementations. This was
just broken in SLURM while adding new environment variables
for BlueGene/Q sub-block jobs.

Patch from Jon Bringhurst, LANL.

Add abstracts to the Slurm User Group Meeting 2011 agenda

Add job id to the information loggedn when a REQUEST_SPANK_ENVIRONMENT
RPC is issued. Patch from Andriy Grytsenko.

Conflicts:
	src/slurmd/slurmd/slurmd.c

Get slurmd -f option working. Patch from Andriy Grytsenko.