Merge pull request ceph#61594 from adk3798/test-nfs-task-cluster-purge-fixup

mgr/cephadm: continue in nfs service purge if grace file is already deleted

Reviewed-by: John Mulligan <[email protected]>

Author: adk3798

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

@@ -14,7 +14,7 @@
 from ceph.deployment.service_spec import ServiceSpec, NFSServiceSpec
 from .service_registry import register_cephadm_service
 
-from orchestrator import DaemonDescription
+from orchestrator import DaemonDescription, OrchestratorError
 
 from cephadm.services.cephadmservice import AuthEntity, CephadmDaemonDeploySpec, CephService
 
@@ -321,12 +321,24 @@ def purge(self, service_name: str) -> None:
             '--namespace', cast(str, spec.service_id),
             'rm', 'grace',
         ]
-        subprocess.run(
-            cmd,
-            stdout=subprocess.PIPE,
-            stderr=subprocess.PIPE,
-            timeout=10
-        )
+        try:
+            result = subprocess.run(
+                cmd,
+                stdout=subprocess.PIPE,
+                stderr=subprocess.PIPE,
+                timeout=10
+            )
+        except Exception as e:
+            err_msg = f'Got unexpected exception trying to remove ganesha grace file for nfs.{spec.service_id} service: {str(e)}'
+            self.mgr.log.warning(err_msg)
+            raise OrchestratorError(err_msg)
+        if result.returncode:
+            if "No such file" in result.stderr.decode('utf-8'):
+                logger.info(f'Grace file for nfs.{spec.service_id} already deleted')
+            else:
+                err_msg = f'Failed to remove ganesha grace file for nfs.{spec.service_id} service: {result.stderr.decode("utf-8")}'
+                self.mgr.log.warning(err_msg)
+                raise OrchestratorError(err_msg)
 
     def _haproxy_hosts(self) -> List[str]:
         # NB: Ideally, we would limit the list to IPs on hosts running