LU-8324 hsm: prioritize one RESTORE once in a while

LU-8324 hsm: prioritize one RESTORE once in a while

Currently, HSM requests are run in the order they are submitted to
the coordinator. This has the undesirable side effect that the more
interactive requests (RESTORE / REMOVE) can take quite a while to be
run when a huge batch of ARCHIVE requests are already queued.

This patch is not a clean fix to LU-8324, it is merely an attempt at
making things bearable while a proper solution is being developped.

Test-Parameters: trivial testlist=sanity-hsm
Signed-off-by: Quentin Bouget <quentin.bouget@cea.fr>
Change-Id: I8697ffae19b28f31901d9e61cce55b40f848fb51
Reviewed-on: https://review.whamcloud.com/31723
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>