orch: disk replacement enhancement

This introduces a new `ceph orch device replace` command in order to
improve the user experience when it comes to replacing the underlying
device of an OSD.

Fixes: https://tracker.ceph.com/issues/68456

Signed-off-by: Guillaume Abrioux <gabrioux@ibm.com>

2 files changed, 33 insertions, 2 deletions

diff --git a/src/pybind/mgr/orchestrator/_interface.py b/src/pybind/mgr/orchestrator/_interface.py
index 82a8c13a9c1..c05332df59a 100644
--- a/src/pybind/mgr/orchestrator/_interface.py
+++ b/src/pybind/mgr/orchestrator/_interface.py
@@ -520,6 +520,15 @@ class Orchestrator(object):
         """
         raise NotImplementedError()
 
+    def replace_device(self,
+                       hostname: str,
+                       device: str,
+                       clear: bool = False,
+                       yes_i_really_mean_it: bool = False) -> OrchResult:
+        """Perform all required operations in order to replace a device.
+        """
+        raise NotImplementedError()
+
     def get_inventory(self, host_filter: Optional['InventoryFilter'] = None, refresh: bool = False) -> OrchResult[List['InventoryHost']]:
         """
         Returns something that was created by `ceph-volume inventory`.
@@ -699,12 +708,18 @@ class Orchestrator(object):
 
     def remove_osds(self, osd_ids: List[str],
                     replace: bool = False,
+                    replace_block: bool = False,
+                    replace_db: bool = False,
+                    replace_wal: bool = False,
                     force: bool = False,
                     zap: bool = False,
                     no_destroy: bool = False) -> OrchResult[str]:
         """
         :param osd_ids: list of OSD IDs
         :param replace: marks the OSD as being destroyed. See :ref:`orchestrator-osd-replace`
+        :param replace_block: marks the corresponding block device as being replaced.
+        :param replace_db: marks the corresponding db device as being replaced.
+        :param replace_wal: marks the corresponding wal device as being replaced.
         :param force: Forces the OSD removal process without waiting for the data to be drained first.
         :param zap: Zap/Erase all devices associated with the OSDs (DESTROYS DATA)
         :param no_destroy: Do not destroy associated VGs/LVs with the OSD.
diff --git a/src/pybind/mgr/orchestrator/module.py b/src/pybind/mgr/orchestrator/module.py
index be0096bb2d9..7dd8c95af52 100644
--- a/src/pybind/mgr/orchestrator/module.py
+++ b/src/pybind/mgr/orchestrator/module.py
@@ -818,6 +818,21 @@ class OrchestratorCli(OrchestratorClientMixin, MgrModule,
             return HandleCommandResult(stdout=completion.result_str())
         return HandleCommandResult(stdout=completion.result_str().split('.')[0])
 
+    @_cli_read_command('orch device replace')
+    def _replace_device(self,
+                        hostname: str,
+                        device: str,
+                        clear: bool = False,
+                        yes_i_really_mean_it: bool = False) -> HandleCommandResult:
+        """Perform all required operations in order to replace a device.
+        """
+        completion = self.replace_device(hostname=hostname,
+                                         device=device,
+                                         clear=clear,
+                                         yes_i_really_mean_it=yes_i_really_mean_it)
+        raise_if_exception(completion)
+        return HandleCommandResult(stdout=completion.result_str())
+
     @_cli_read_command('orch device ls')
     def _list_devices(self,
                       hostname: Optional[List[str]] = None,
@@ -1415,8 +1430,9 @@ Usage:
                       zap: bool = False,
                       no_destroy: bool = False) -> HandleCommandResult:
         """Remove OSD daemons"""
-        completion = self.remove_osds(osd_id, replace=replace, force=force,
-                                      zap=zap, no_destroy=no_destroy)
+        completion = self.remove_osds(osd_id,
+                                      replace=replace,
+                                      force=force, zap=zap, no_destroy=no_destroy)
         raise_if_exception(completion)
         return HandleCommandResult(stdout=completion.result_str())