Merge pull request #14506 from smithfarm/wip-19119-kraken

kraken: doc: PendingReleaseNotes: warning about 'osd rm ...' and #13733 

Reviewed-by: Sage Weil <sage@redhat.com>

PendingReleaseNotes

@@ -1,17 +1,23 @@
-11.1.1
+11.2.1
 ------
 
- * Calculation of recovery priorities has been updated.
-   This could lead to unintuitive recovery prioritization
-   during cluster upgrade. In case of such recovery, OSDs
-   in old version would operate on different priority ranges
-   than new ones. Once upgraded, cluster will operate on
-   consistent values.
-
 * In previous versions, if a client sent an op to the wrong OSD, the OSD
   would reply with ENXIO.  The rationale here is that the client or OSD is
   clearly buggy and we want to surface the error as clearly as possible.
   We now only send the ENXIO reply if the osd_enxio_on_misdirected_op option
   is enabled (it's off by default).  This means that a VM using librbd that
   previously would have gotten an EIO and gone read-only will now see a
   blocked/hung IO instead.
+
+* There was a bug introduced in Jewel (#19119) that broke the mapping behavior
+  when an "out" OSD that still existed in the CRUSH map was removed with 'osd rm'.
+  This could result in 'misdirected op' and other errors.  The bug is now fixed,
+  but the fix itself introduces the same risk because the behavior may vary between
+  clients and OSDs.  To avoid problems, please ensure that all OSDs are removed
+  from the CRUSH map before deleting them.  That is, be sure to do::
+
+     ceph osd crush rm osd.123
+
+  before::
+
+     ceph osd rm osd.123