Document that hole_birth is effectively useless

The first sentence of this commit comes from the wiki, and was
originally written by:
Rich Ercolani <rincebrain@gmail.com>
with changes by:
Tom Caputi <tcaputi@datto.com>

Reviewed-by: Tom Caputi <tcaputi@datto.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Richard Laager <rlaager@wiktel.com>
Closes #8641
Closes #8642

diff --git a/man/man5/zpool-features.5 b/man/man5/zpool-features.5
index 1df7f0f76d41bc525e13dc84245cc026ed27d78e..8893d151d1f088ab3924bac3f76248dc057e3335 100644 (file)
--- a/man/man5/zpool-features.5
+++ b/man/man5/zpool-features.5
@@ -383,6 +383,14 @@ READ\-ONLY COMPATIBLE      no
 DEPENDENCIES   enabled_txg
 .TE
 
+This feature has/had bugs, the result of which is that, if you do a
+\fBzfs send -i\fR (or \fB-R\fR, since it uses \fB-i\fR) from an affected
+dataset, the receiver will not see any checksum or other errors, but the
+resulting destination snapshot will not match the source.  Its use by
+\fBzfs send -i\fR has been disabled by default.  See the
+\fBsend_holes_without_birth_time\fR module parameter in
+zfs-module-parameters(5).
+
 This feature improves performance of incremental sends (\fBzfs send -i\fR)
 and receives for objects with many holes. The most common case of
 hole-filled objects is zvols.