mgr/cephadm: fix 'cephadm osd activate' on existing osd devices

qa/suites/orch/cephadm/osds/2-ops/rmdir-reactivate.yaml

@@ -0,0 +1,20 @@
+tasks:
+- cephadm.shell:
+    host.a:
+      - |
+        set -e
+        set -x
+        ceph orch ps
+        HOST=$(hostname -s)
+        OSD=$(ceph orch ps $HOST | grep osd | head -n 1 | awk '{print $1}')
+        echo "host $HOST, osd $OSD"
+        ceph orch daemon stop $OSD
+        while ceph orch ps | grep $OSD | grep running ; do sleep 5 ; done
+        ceph auth export $OSD > k
+        ceph orch daemon rm $OSD --force
+        ceph orch ps --refresh
+        while ceph orch ps | grep $OSD ; do sleep 5 ; done
+        ceph auth add $OSD -i k
+        ceph cephadm osd activate $HOST
+        while ! ceph orch ps | grep $OSD | grep running ; do sleep 5 ; done
+- cephadm.healthy:

src/pybind/mgr/cephadm/inventory.py

@@ -825,9 +825,9 @@ def get_daemon(self, daemon_name: str, host: Optional[str] = None) -> orchestrat
 
         raise orchestrator.OrchestratorError(f'Unable to find {daemon_name} daemon(s)')
 
-    def has_daemon(self, daemon_name: str) -> bool:
+    def has_daemon(self, daemon_name: str, host: Optional[str] = None) -> bool:
         try:
-            self.get_daemon(daemon_name)
+            self.get_daemon(daemon_name, host)
         except orchestrator.OrchestratorError:
             return False
         return True

src/pybind/mgr/cephadm/module.py

@@ -2125,7 +2125,7 @@ def get_osd_uuid_map(self, only_up=False):
             osd_id = o.get('osd')
             if osd_id is None:
                 raise OrchestratorError("Could not retrieve osd_id from osd_map")
-            if not only_up or (o['up_from'] > 0):
+            if not only_up:
                 r[str(osd_id)] = o.get('uuid', '')
         return r
 

src/pybind/mgr/cephadm/services/osd.py

@@ -125,6 +125,10 @@ async def deploy_osd_daemons_for_existing_osds(self, host: str, service_name: st
                 if osd_id in before_osd_uuid_map and osd_id not in replace_osd_ids:
                     # if it exists but is part of the replacement operation, don't skip
                     continue
+                if self.mgr.cache.has_daemon(f'osd.{osd_id}', host):
+                    # cephadm daemon instance already exists
+                    logger.debug(f'osd id {osd_id} daemon already exists')
+                    continue
                 if osd_id not in osd_uuid_map:
                     logger.debug('osd id {} does not exist in cluster'.format(osd_id))
                     continue
@@ -162,6 +166,10 @@ async def deploy_osd_daemons_for_existing_osds(self, host: str, service_name: st
             if osd_id in before_osd_uuid_map and osd_id not in replace_osd_ids:
                 # if it exists but is part of the replacement operation, don't skip
                 continue
+            if self.mgr.cache.has_daemon(f'osd.{osd_id}', host):
+                # cephadm daemon instance already exists
+                logger.debug(f'osd id {osd_id} daemon already exists')
+                continue
             if osd_id not in osd_uuid_map:
                 logger.debug('osd id {} does not exist in cluster'.format(osd_id))
                 continue