See previous unreverted commit.

This reverts commit 2ec92d36a8ad7184897c9a322ba2d9978d2ccdbd.

This is an example of how to do it. The problem is that select_jobinfo
on the job_desc is packed using working_cluster's->plugin_id.
job_desc's->select_jobinfo is only used by bluegene and alps code which
will eventually go away.

sbatch will choose the federation or individual cluster with the fast
startime. A willrun to a federation will return the fastest start time
of all clusters in a federation.

with the sbatch -M<clusters> option. A fed will run will the fastest
time of all siblings in a federation.

Was moved in 5cfc577a

More to come. This sets up the controller side.

It could have state saved a null federation.

instead of one at a time. Saves ~20 seconds.

If all of the clusters get updated at the same time, the could get in a
state where they are waiting on each other to respond and will
eventually timeout and will then will reconnect. By spacing out the
clusters being added this helps prevent them from talking to everyone at
the same time.

get_next_job_id() didn't take into consideration job_ids that already be
taken by other jobs like set_job_id() did.

Send the existing packed buffer that has the job_desc to the sibling.
The sibling will unpack it on the other side. This prevents having to
pack the job_desc for each willrun/allocation to each sibling.

squeue --fedtrack

squeue long options: fedorigin, fedoriginraw, fedsiblings and
fedsiblingsraw.

Also make strings of siblings for passing back to the api.

For the fed_mgr, instead of packing up the job_desc again for will_runs
and allocations to each sibling just send the buffer that came in.

Since job_allocate will modify the received job_desc_msg_t, this also
makes it easy to make a copy of the received job_desc_msg_t to do a
willrun on the receiving cluster before allocating resources.

This option specifies minimum characteristics of the compute nodes
  which should be considered for use, not the resource allocation
  size.
bug 3118

Create separate check_hosts_contiguous procedure in globals and use it
for both test1.83 and test15.21.

Bug 3006.

Fix bug that was clearing MAINT mode on nodes scheduled for reboot (bug
    introduced in version 16.05.5 to address bug in overlapping reservations,
    commit 5eee1d28). Note that a node's
    MAINT flag is used for both a requested reboot and maintenance reservation.
    What I'd like to do is add a new node state flag to differenciate
    between these two cases, but that involves some significant changes
    that could introduce instability, so it will be defered to version
    17.02
bug 3210

Correct/expand description of NODE_STATE_FLAG