For k8s, use provisioner-managed volumes when an absolute host path is not specified. (#741)

In kind, when we bind-mount a host directory it is first mounted into the kind container at /mnt, then into the pod at the desired location.

We accidentally picked this up for full-blown k8s, and were creating volumes at /mnt.  This changes the behavior for both kind and regular k8s so that bind mounts are only allowed if a fully-qualified path is specified.  If no path is specified at all, a default storageClass is assumed to be present, and the volume managed by a provisioner.

Eg, for kind, the default provisioner is: https://github.com/rancher/local-path-provisioner

```
stack: test
deploy-to: k8s-kind
config:
  test-variable-1: test-value-1
network:
  ports:
    test:
     - '80'
volumes:
  # this will be bind-mounted to a host-path
  test-data-bind: /srv/data
  # this will be managed by the k8s node
  test-data-auto:
configmaps:
  test-config: ./configmap/test-config
```

Reviewed-on: #741
Co-authored-by: Thomas E Lackey <telackey@bozemanpass.com>
Co-committed-by: Thomas E Lackey <telackey@bozemanpass.com>

stack_orchestrator/deploy/k8s/cluster_info.py

@@ -21,7 +21,7 @@ from typing import Any, List, Set
 from stack_orchestrator.opts import opts
 from stack_orchestrator.util import env_var_map_from_file
 from stack_orchestrator.deploy.k8s.helpers import named_volumes_from_pod_files, volume_mounts_for_service, volumes_for_pod_files
-from stack_orchestrator.deploy.k8s.helpers import get_node_pv_mount_path
+from stack_orchestrator.deploy.k8s.helpers import get_kind_pv_bind_mount_path
 from stack_orchestrator.deploy.k8s.helpers import envs_from_environment_variables_map, envs_from_compose_file, merge_envs
 from stack_orchestrator.deploy.deploy_util import parsed_pod_files_map_from_file_names, images_for_deployment
 from stack_orchestrator.deploy.deploy_types import DeployEnvVars
@@ -171,21 +171,33 @@ class ClusterInfo:
             print(f"Spec Volumes: {spec_volumes}")
             print(f"Named Volumes: {named_volumes}")
             print(f"Resources: {resources}")
-        for volume_name in spec_volumes:
+        for volume_name, volume_path in spec_volumes.items():
             if volume_name not in named_volumes:
                 if opts.o.debug:
                     print(f"{volume_name} not in pod files")
                 continue
+
+            labels = {
+                "app": self.app_name,
+                "volume-label": f"{self.app_name}-{volume_name}"
+            }
+            if volume_path:
+                storage_class_name = "manual"
+                k8s_volume_name = f"{self.app_name}-{volume_name}"
+            else:
+                # These will be auto-assigned.
+                storage_class_name = None
+                k8s_volume_name = None
+
             spec = client.V1PersistentVolumeClaimSpec(
                 access_modes=["ReadWriteOnce"],
-                storage_class_name="manual",
+                storage_class_name=storage_class_name,
                 resources=to_k8s_resource_requirements(resources),
-                volume_name=f"{self.app_name}-{volume_name}"
+                volume_name=k8s_volume_name
             )
             pvc = client.V1PersistentVolumeClaim(
-                metadata=client.V1ObjectMeta(name=f"{self.app_name}-{volume_name}",
-                                             labels={"volume-label": f"{self.app_name}-{volume_name}"}),
-                spec=spec,
+                metadata=client.V1ObjectMeta(name=f"{self.app_name}-{volume_name}", labels=labels),
+                spec=spec
             )
             result.append(pvc)
         return result
@@ -226,16 +238,32 @@ class ClusterInfo:
         resources = self.spec.get_volume_resources()
         if not resources:
             resources = DEFAULT_VOLUME_RESOURCES
-        for volume_name in spec_volumes:
+        for volume_name, volume_path in spec_volumes.items():
+            # We only need to create a volume if it is fully qualified HostPath.
+            # Otherwise, we create the PVC and expect the node to allocate the volume for us.
+            if not volume_path:
+                if opts.o.debug:
+                    print(f"{volume_name} does not require an explicit PersistentVolume, since it is not a bind-mount.")
+                continue
+
             if volume_name not in named_volumes:
                 if opts.o.debug:
                     print(f"{volume_name} not in pod files")
                 continue
+
+            if not os.path.isabs(volume_path):
+                print(f"WARNING: {volume_name}:{volume_path} is not absolute, cannot bind volume.")
+                continue
+
+            if self.spec.is_kind_deployment():
+                host_path = client.V1HostPathVolumeSource(path=get_kind_pv_bind_mount_path(volume_name))
+            else:
+                host_path = client.V1HostPathVolumeSource(path=volume_path)
             spec = client.V1PersistentVolumeSpec(
                 storage_class_name="manual",
                 access_modes=["ReadWriteOnce"],
                 capacity=to_k8s_resource_requirements(resources).requests,
-                host_path=client.V1HostPathVolumeSource(path=get_node_pv_mount_path(volume_name))
+                host_path=host_path
             )
             pv = client.V1PersistentVolume(
                 metadata=client.V1ObjectMeta(name=f"{self.app_name}-{volume_name}",

stack_orchestrator/deploy/k8s/deploy_k8s.py

@@ -88,6 +88,16 @@ class K8sDeployer(Deployer):
             if opts.o.debug:
                 print(f"Sending this pv: {pv}")
             if not opts.o.dry_run:
+                try:
+                    pv_resp = self.core_api.read_persistent_volume(name=pv.metadata.name)
+                    if pv_resp:
+                        if opts.o.debug:
+                            print("PVs already present:")
+                            print(f"{pv_resp}")
+                        continue
+                except:  # noqa: E722
+                    pass
+
                 pv_resp = self.core_api.create_persistent_volume(body=pv)
                 if opts.o.debug:
                     print("PVs created:")
@@ -100,6 +110,17 @@ class K8sDeployer(Deployer):
                 print(f"Sending this pvc: {pvc}")
 
             if not opts.o.dry_run:
+                try:
+                    pvc_resp = self.core_api.read_namespaced_persistent_volume_claim(
+                        name=pvc.metadata.name, namespace=self.k8s_namespace)
+                    if pvc_resp:
+                        if opts.o.debug:
+                            print("PVCs already present:")
+                            print(f"{pvc_resp}")
+                        continue
+                except:  # noqa: E722
+                    pass
+
                 pvc_resp = self.core_api.create_namespaced_persistent_volume_claim(body=pvc, namespace=self.k8s_namespace)
                 if opts.o.debug:
                     print("PVCs created:")
@@ -181,33 +202,35 @@ class K8sDeployer(Deployer):
     def down(self, timeout, volumes):  # noqa: C901
         self.connect_api()
         # Delete the k8s objects
-        # Create the host-path-mounted PVs for this deployment
-        pvs = self.cluster_info.get_pvs()
-        for pv in pvs:
-            if opts.o.debug:
-                print(f"Deleting this pv: {pv}")
-            try:
-                pv_resp = self.core_api.delete_persistent_volume(name=pv.metadata.name)
-                if opts.o.debug:
-                    print("PV deleted:")
-                    print(f"{pv_resp}")
-            except client.exceptions.ApiException as e:
-                _check_delete_exception(e)
 
-        # Figure out the PVCs for this deployment
-        pvcs = self.cluster_info.get_pvcs()
-        for pvc in pvcs:
-            if opts.o.debug:
-                print(f"Deleting this pvc: {pvc}")
-            try:
-                pvc_resp = self.core_api.delete_namespaced_persistent_volume_claim(
-                    name=pvc.metadata.name, namespace=self.k8s_namespace
-                )
+        if volumes:
+            # Create the host-path-mounted PVs for this deployment
+            pvs = self.cluster_info.get_pvs()
+            for pv in pvs:
                 if opts.o.debug:
-                    print("PVCs deleted:")
-                    print(f"{pvc_resp}")
-            except client.exceptions.ApiException as e:
-                _check_delete_exception(e)
+                    print(f"Deleting this pv: {pv}")
+                try:
+                    pv_resp = self.core_api.delete_persistent_volume(name=pv.metadata.name)
+                    if opts.o.debug:
+                        print("PV deleted:")
+                        print(f"{pv_resp}")
+                except client.exceptions.ApiException as e:
+                    _check_delete_exception(e)
+
+            # Figure out the PVCs for this deployment
+            pvcs = self.cluster_info.get_pvcs()
+            for pvc in pvcs:
+                if opts.o.debug:
+                    print(f"Deleting this pvc: {pvc}")
+                try:
+                    pvc_resp = self.core_api.delete_namespaced_persistent_volume_claim(
+                        name=pvc.metadata.name, namespace=self.k8s_namespace
+                    )
+                    if opts.o.debug:
+                        print("PVCs deleted:")
+                        print(f"{pvc_resp}")
+                except client.exceptions.ApiException as e:
+                    _check_delete_exception(e)
 
         # Figure out the ConfigMaps for this deployment
         cfg_maps = self.cluster_info.get_configmaps()

stack_orchestrator/deploy/k8s/helpers.py

@@ -81,7 +81,7 @@ def named_volumes_from_pod_files(parsed_pod_files):
     return named_volumes
 
 
-def get_node_pv_mount_path(volume_name: str):
+def get_kind_pv_bind_mount_path(volume_name: str):
     return f"/mnt/{volume_name}"
 
 
@@ -106,11 +106,14 @@ def volume_mounts_for_service(parsed_pod_files, service):
                             mount_path = mount_split[1]
                             mount_options = mount_split[2] if len(mount_split) == 3 else None
                             if opts.o.debug:
-                                print(f"volumne_name: {volume_name}")
+                                print(f"volume_name: {volume_name}")
                                 print(f"mount path: {mount_path}")
                                 print(f"mount options: {mount_options}")
                             volume_device = client.V1VolumeMount(
-                                mount_path=mount_path, name=volume_name, read_only="ro" == mount_options)
+                                mount_path=mount_path,
+                                name=volume_name,
+                                read_only="ro" == mount_options
+                            )
                             result.append(volume_device)
     return result
 
@@ -133,18 +136,8 @@ def volumes_for_pod_files(parsed_pod_files, spec, app_name):
     return result
 
 
-def _get_host_paths_for_volumes(parsed_pod_files):
-    result = {}
-    for pod in parsed_pod_files:
-        parsed_pod_file = parsed_pod_files[pod]
-        if "volumes" in parsed_pod_file:
-            volumes = parsed_pod_file["volumes"]
-            for volume_name in volumes.keys():
-                volume_definition = volumes[volume_name]
-                if volume_definition and "driver_opts" in volume_definition:
-                    host_path = volume_definition["driver_opts"]["device"]
-                    result[volume_name] = host_path
-    return result
+def _get_host_paths_for_volumes(deployment_context):
+    return deployment_context.spec.get_volumes()
 
 
 def _make_absolute_host_path(data_mount_path: Path, deployment_dir: Path) -> Path:
@@ -152,12 +145,12 @@ def _make_absolute_host_path(data_mount_path: Path, deployment_dir: Path) -> Pat
         return data_mount_path
     else:
         # Python Path voodo that looks pretty odd:
-        return Path.cwd().joinpath(deployment_dir.joinpath("compose").joinpath(data_mount_path)).resolve()
+        return Path.cwd().joinpath(deployment_dir.joinpath(data_mount_path)).resolve()
 
 
 def _generate_kind_mounts(parsed_pod_files, deployment_dir, deployment_context):
     volume_definitions = []
-    volume_host_path_map = _get_host_paths_for_volumes(parsed_pod_files)
+    volume_host_path_map = _get_host_paths_for_volumes(deployment_context)
     # Note these paths are relative to the location of the pod files (at present)
     # So we need to fix up to make them correct and absolute because kind assumes
     # relative to the cwd.
@@ -177,14 +170,15 @@ def _generate_kind_mounts(parsed_pod_files, deployment_dir, deployment_context):
                         volume_name = mount_split[0]
                         mount_path = mount_split[1]
                         if opts.o.debug:
-                            print(f"volumne_name: {volume_name}")
+                            print(f"volume_name: {volume_name}")
                             print(f"map: {volume_host_path_map}")
                             print(f"mount path: {mount_path}")
                         if volume_name not in deployment_context.spec.get_configmaps():
-                            volume_definitions.append(
-                                f"  - hostPath: {_make_absolute_host_path(volume_host_path_map[volume_name], deployment_dir)}\n"
-                                f"    containerPath: {get_node_pv_mount_path(volume_name)}\n"
-                            )
+                            if volume_host_path_map[volume_name]:
+                                volume_definitions.append(
+                                    f"  - hostPath: {_make_absolute_host_path(volume_host_path_map[volume_name], deployment_dir)}\n"
+                                    f"    containerPath: {get_kind_pv_bind_mount_path(volume_name)}\n"
+                                )
     return (
         "" if len(volume_definitions) == 0 else (
             "  extraMounts:\n"