osd/scrub: remove the 'has_deep_errors' scheduling flag

osd/scrub: remove the 'has_deep_errors' scheduling flag

Following https://github.com/ceph/ceph/pull/59942
("osd/scrub: separate shallow vs deep errors storage"),
the ScrubStore now holds two separate error databases,
one for the shallow errors and one for the deep errors.

There is no longer a need to prevent (or warn about)
shallow scrubs being performed when deep errors are present.

Signed-off-by: Ronen Friedman <rfriedma@redhat.com>