LU-2147 quota: several fixes to reintegration procedure

LU-2147 quota: several fixes to reintegration procedure

This patch gathers several fixes/improvements to the quota
reintegration procedure:
- do not set rq_no_resend & rq_no_delay for IT_QUOTA_CONN to have
  the reintegration thread waiting instead of stopping/starting
  the reint thread until the master is available
- add procfs tunable to force reintegration, it can be useful for
  testing, but also for fixing things at customer site when a bug
  was hit during reintegration.
- when transferring indexes, the per-page header isn't swabbed
- on index transfer, the hash value isn't sent any more (unlike on
  orion_quota) since we now use II_FL_NOHASH. As a consequence,
  qsd_reint_entries() shouldn't take the hash size into account
  when parsing a page container key/record pairs.

This patch also:
- quiets many common messages which aren't real errors and shouldn't
  make it to the console
- fixes a bug in qmt_adjust_edquot() which does not check correctly
  whether the revoke timeout has elapsed
- changes test_6 to use a larger quota limit to avoid edquot flag
  to be set by the QMT and cause the test to fail.
- removes temporary code from setup_quota() in t-f now that the
  new quota code is fully landed.
- re-enables ost-pool tests 23a & 23b

Signed-off-by: Johann Lombardi <johann.lombardi@intel.com>
Change-Id: I9af9a025faa1ef173810df647b93307e2139c6f9
Reviewed-on: http://review.whamcloud.com/4253
Tested-by: Hudson
Reviewed-by: Niu Yawei <niu@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>