osd/scrub: tag replica scrub messages to identify stale events

osd/scrub: tag replica scrub messages to identify stale events

A lot can happen while a replica is waiting for the backend
to collect the map data. The Primary might, for example, abort
the scrub and start a new one (following no-scrub commands).

The 'token' introduced here tags 'replica scrub resched'
messages with an index value that is modified on each 'release
scrub resources' request from the Primary.

Fixes: https://tracker.ceph.com/issues/52012
Signed-off-by: Ronen Friedman <rfriedma@redhat.com>