Bump version to 0.9.94

Adds the same validations as a volume add or resize to volume clones, to
ensure there is enough free space for them.

.version

@@ -1 +1 @@
-0.9.90
+0.9.94

CHANGELOG.md

@@ -1,5 +1,32 @@
 ## PVC Changelog
 
+###### [v0.9.94](https://github.com/parallelvirtualcluster/pvc/releases/tag/v0.9.94)
+
+  * [CLI Client] Fixes an incorrect ordering issue with autobackup summary emails
+  * [API Daemon/CLI Client] Adds an additional safety check for 80% cluster fullness when doing volume adds or resizes
+  * [API Daemon/CLI Client] Adds safety checks to volume clones as well
+  * [API Daemon] Fixes a few remaining memory bugs for stopped/disabled VMs
+
+###### [v0.9.93](https://github.com/parallelvirtualcluster/pvc/releases/tag/v0.9.93)
+
+  * [API Daemon] Fixes a bug where stuck zkhandler threads were not cleaned up on error
+
+###### [v0.9.92](https://github.com/parallelvirtualcluster/pvc/releases/tag/v0.9.92)
+
+  * [CLI Client] Adds the new restore state to the colours list for VM status
+  * [API Daemon] Fixes an incorrect variable assignment
+  * [Provisioner] Improves the error handling of various steps in the debootstrap and rinse example scripts
+  * [CLI Client] Fixes two bugs around missing keys that were added recently (uses get() instead direct dictionary refs)
+  * [CLI Client] Improves API error handling via GET retries (x3) and better server status code handling
+
+###### [v0.9.91](https://github.com/parallelvirtualcluster/pvc/releases/tag/v0.9.91)
+
+  * [Client CLI] Fixes a bug and improves output during cluster task events.
+  * [Client CLI] Improves the output of the task list display.
+  * [Provisioner] Fixes some missing cloud-init modules in the default debootstrap script.
+  * [Client CLI] Fixes a bug with a missing argument to the vm_define helper function.
+  * [All] Fixes inconsistent package find + rm commands to avoid errors in dpkg.
+
 ###### [v0.9.90](https://github.com/parallelvirtualcluster/pvc/releases/tag/v0.9.90)
 
   * [Client CLI/API Daemon] Adds additional backup metainfo and an emailed report option to autobackups.

api-daemon/provisioner/examples/script/3-debootstrap.py

@@ -150,6 +150,10 @@
 from daemon_lib.vmbuilder import VMBuilder
 
 
+# These are some global variables used below
+default_root_password = "test123"
+
+
 # The VMBuilderScript class must be named as such, and extend VMBuilder.
 class VMBuilderScript(VMBuilder):
     def setup(self):
@@ -498,11 +502,15 @@ class VMBuilderScript(VMBuilder):
         ret = os.system(
             f"debootstrap --include={','.join(deb_packages)} {deb_release} {temp_dir} {deb_mirror}"
         )
+        ret = int(ret >> 8)
         if ret > 0:
-            self.fail("Failed to run debootstrap")
+            self.fail(f"Debootstrap failed with exit code {ret}")
 
         # Bind mount the devfs so we can grub-install later
-        os.system("mount --bind /dev {}/dev".format(temp_dir))
+        ret = os.system("mount --bind /dev {}/dev".format(temp_dir))
+        ret = int(ret >> 8)
+        if ret > 0:
+            self.fail(f"/dev bind mount failed with exit code {ret}")
 
         # Create an fstab entry for each volume
         fstab_file = "{}/etc/fstab".format(temp_dir)
@@ -589,11 +597,13 @@ After=multi-user.target
                  - migrator
                  - bootcmd
                  - write-files
+                 - growpart
                  - resizefs
                  - set_hostname
                  - update_hostname
                  - update_etc_hosts
                  - ca-certs
+                 - users-groups
                  - ssh
                 
                 cloud_config_modules:
@@ -686,23 +696,36 @@ GRUB_DISABLE_LINUX_UUID=false
         # Do some tasks inside the chroot using the provided context manager
         with chroot(temp_dir):
             # Install and update GRUB
-            os.system(
+            ret = os.system(
                 "grub-install --force /dev/rbd/{}/{}_{}".format(
                     root_volume["pool"], vm_name, root_volume["disk_id"]
                 )
             )
-            os.system("update-grub")
+            ret = int(ret >> 8)
+            if ret > 0:
+                self.fail(f"GRUB install failed with exit code {ret}")
+
+            ret = os.system("update-grub")
+            ret = int(ret >> 8)
+            if ret > 0:
+                self.fail(f"GRUB update failed with exit code {ret}")
 
             # Set a really dumb root password so the VM can be debugged
             # EITHER CHANGE THIS YOURSELF, here or in Userdata, or run something after install
             # to change the root password: don't leave it like this on an Internet-facing machine!
-            os.system("echo root:test123 | chpasswd")
+            ret = os.system(f"echo root:{default_root_password} | chpasswd")
+            ret = int(ret >> 8)
+            if ret > 0:
+                self.fail(f"Root password change failed with exit code {ret}")
 
             # Enable cloud-init target on (first) boot
             # Your user-data should handle this and disable it once done, or things get messy.
             # That cloud-init won't run without this hack seems like a bug... but even the official
             # Debian cloud images are affected, so who knows.
-            os.system("systemctl enable cloud-init.target")
+            ret = os.system("systemctl enable cloud-init.target")
+            ret = int(ret >> 8)
+            if ret > 0:
+                self.fail(f"Enable of cloud-init failed with exit code {ret}")
 
     def cleanup(self):
         """
@@ -727,7 +750,7 @@ GRUB_DISABLE_LINUX_UUID=false
         temp_dir = "/tmp/target"
 
         # Unmount the bound devfs
-        os.system("umount {}/dev".format(temp_dir))
+        os.system("umount -f {}/dev".format(temp_dir))
 
         # Use this construct for reversing the list, as the normal reverse() messes with the list
         for volume in list(reversed(self.vm_data["volumes"])):
@@ -744,7 +767,7 @@ GRUB_DISABLE_LINUX_UUID=false
             ):
                 # Unmount filesystem
                 retcode, stdout, stderr = pvc_common.run_os_command(
-                    f"umount {mount_path}"
+                    f"umount -f {mount_path}"
                 )
                 if retcode:
                     self.log_err(

api-daemon/provisioner/examples/script/4-rinse.py

@@ -150,6 +150,11 @@
 from daemon_lib.vmbuilder import VMBuilder
 
 
+# These are some global variables used below
+default_root_password = "test123"
+default_local_time = "UTC"
+
+
 # The VMBuilderScript class must be named as such, and extend VMBuilder.
 class VMBuilderScript(VMBuilder):
     def setup(self):
@@ -524,13 +529,23 @@ class VMBuilderScript(VMBuilder):
         ret = os.system(
             f"rinse --arch {rinse_architecture} --directory {temporary_directory} --distribution {rinse_release} --cache-dir {rinse_cache} --add-pkg-list /tmp/addpkg --verbose {mirror_arg}"
         )
+        ret = int(ret >> 8)
         if ret > 0:
-            self.fail("Failed to run rinse")
+            self.fail(f"Rinse failed with exit code {ret}")
 
         # Bind mount the devfs, sysfs, and procfs so we can grub-install later
-        os.system("mount --bind /dev {}/dev".format(temporary_directory))
-        os.system("mount --bind /sys {}/sys".format(temporary_directory))
-        os.system("mount --bind /proc {}/proc".format(temporary_directory))
+        ret = os.system("mount --bind /dev {}/dev".format(temporary_directory))
+        ret = int(ret >> 8)
+        if ret > 0:
+            self.fail(f"/dev bind mount failed with exit code {ret}")
+        ret = os.system("mount --bind /sys {}/sys".format(temporary_directory))
+        ret = int(ret >> 8)
+        if ret > 0:
+            self.fail(f"/sys bind mount failed with exit code {ret}")
+        ret = os.system("mount --bind /proc {}/proc".format(temporary_directory))
+        ret = int(ret >> 8)
+        if ret > 0:
+            self.fail(f"/proc bind mount failed with exit code {ret}")
 
         # Create an fstab entry for each volume
         fstab_file = "{}/etc/fstab".format(temporary_directory)
@@ -642,41 +657,76 @@ GRUB_SERIAL_COMMAND="serial --speed=115200 --unit=0 --word=8 --parity=no --stop=
         # Do some tasks inside the chroot using the provided context manager
         with chroot(temporary_directory):
             # Fix the broken kernel from rinse by setting a systemd machine ID and running the post scripts
-            os.system("systemd-machine-id-setup")
-            os.system(
+            ret = os.system("systemd-machine-id-setup")
+            ret = int(ret >> 8)
+            if ret > 0:
+                self.fail(f"Machine ID setup failed with exit code {ret}")
+
+            ret = os.system(
                 "rpm -q --scripts kernel-core | grep -A20  'posttrans scriptlet' | tail -n+2 | bash -x"
             )
+            ret = int(ret >> 8)
+            if ret > 0:
+                self.fail(f"RPM kernel reinstall failed with exit code {ret}")
 
             # Install any post packages
-            os.system(f"dnf install -y {' '.join(post_packages)}")
+            if len(post_packages) > 0:
+                ret = os.system(f"dnf install -y {' '.join(post_packages)}")
+                ret = int(ret >> 8)
+                if ret > 0:
+                    self.fail(f"DNF install failed with exit code {ret}")
 
             # Install and update GRUB config
-            os.system(
+            ret = os.system(
                 "grub2-install --force /dev/rbd/{}/{}_{}".format(
                     root_volume["pool"], vm_name, root_volume["disk_id"]
                 )
             )
+            ret = int(ret >> 8)
+            if ret > 0:
+                self.fail(f"GRUB install failed with exit code {ret}")
+
             os.system("grub2-mkconfig -o /boot/grub2/grub.cfg")
+            ret = int(ret >> 8)
+            if ret > 0:
+                self.fail(f"GRUB update failed with exit code {ret}")
 
             # Set a really dumb root password so the VM can be debugged
             # EITHER CHANGE THIS YOURSELF, here or in Userdata, or run something after install
             # to change the root password: don't leave it like this on an Internet-facing machine!
-            os.system("echo root:test123 | chpasswd")
+            ret = os.system(f"echo root:{default_root_password} | chpasswd")
+            ret = int(ret >> 8)
+            if ret > 0:
+                self.fail(f"Root password change failed with exit code {ret}")
 
             # Enable dbus-broker
-            os.system("systemctl enable dbus-broker.service")
+            ret = os.system("systemctl enable dbus-broker.service")
+            ret = int(ret >> 8)
+            if ret > 0:
+                self.fail(f"Enable of dbus-broker failed with exit code {ret}")
 
             # Enable NetworkManager
             os.system("systemctl enable NetworkManager.service")
+            ret = int(ret >> 8)
+            if ret > 0:
+                self.fail(f"Enable of NetworkManager failed with exit code {ret}")
 
             # Enable cloud-init target on (first) boot
             # Your user-data should handle this and disable it once done, or things get messy.
             # That cloud-init won't run without this hack seems like a bug... but even the official
             # Debian cloud images are affected, so who knows.
             os.system("systemctl enable cloud-init.target")
+            ret = int(ret >> 8)
+            if ret > 0:
+                self.fail(f"Enable of cloud-init failed with exit code {ret}")
 
             # Set the timezone to UTC
-            os.system("ln -sf ../usr/share/zoneinfo/UTC /etc/localtime")
+            ret = os.system(
+                f"ln -sf ../usr/share/zoneinfo/{default_local_time} /etc/localtime"
+            )
+            ret = int(ret >> 8)
+            if ret > 0:
+                self.fail(f"Localtime update failed with exit code {ret}")
 
     def cleanup(self):
         """

api-daemon/pvcapid/Daemon.py

@@ -27,7 +27,7 @@ from distutils.util import strtobool as dustrtobool
 import daemon_lib.config as cfg
 
 # Daemon version
-version = "0.9.90"
+version = "0.9.94"
 
 # API version
 API_VERSION = 1.0

api-daemon/pvcapid/flaskapi.py

@@ -5744,6 +5744,10 @@ class API_Storage_Ceph_Volume_Root(Resource):
                 "required": True,
                 "helptext": "A volume size in bytes (B implied or with SI suffix k/M/G/T) must be specified.",
             },
+            {
+                "name": "force",
+                "required": False,
+            },
         ]
     )
     @Authenticator
@@ -5769,6 +5773,12 @@ class API_Storage_Ceph_Volume_Root(Resource):
             type: string
             required: true
             description: The volume size, in bytes (B implied) or with a single-character SI suffix (k/M/G/T)
+          - in: query
+            name: force
+            type: boolean
+            required: false
+            default: flase
+            description: Force action if volume creation would violate 80% full soft cap on the pool
         responses:
           200:
             description: OK
@@ -5785,6 +5795,7 @@ class API_Storage_Ceph_Volume_Root(Resource):
             reqargs.get("pool", None),
             reqargs.get("volume", None),
             reqargs.get("size", None),
+            reqargs.get("force", False),
         )
 
 
@@ -5819,7 +5830,11 @@ class API_Storage_Ceph_Volume_Element(Resource):
                 "name": "size",
                 "required": True,
                 "helptext": "A volume size in bytes (or with k/M/G/T suffix) must be specified.",
-            }
+            },
+            {
+                "name": "force",
+                "required": False,
+            },
         ]
     )
     @Authenticator
@@ -5835,6 +5850,12 @@ class API_Storage_Ceph_Volume_Element(Resource):
             type: string
             required: true
             description: The volume size in bytes (or with a metric suffix, i.e. k/M/G/T)
+          - in: query
+            name: force
+            type: boolean
+            required: false
+            default: flase
+            description: Force action if volume creation would violate 80% full soft cap on the pool
         responses:
           200:
             description: OK
@@ -5852,9 +5873,17 @@ class API_Storage_Ceph_Volume_Element(Resource):
               type: object
               id: Message
         """
-        return api_helper.ceph_volume_add(pool, volume, reqargs.get("size", None))
+        return api_helper.ceph_volume_add(
+            pool, volume, reqargs.get("size", None), reqargs.get("force", False)
+        )
 
-    @RequestParser([{"name": "new_size"}, {"name": "new_name"}])
+    @RequestParser(
+        [
+            {"name": "new_size"},
+            {"name": "new_name"},
+            {"name": "force", "required": False},
+        ]
+    )
     @Authenticator
     def put(self, pool, volume, reqargs):
         """
@@ -5873,6 +5902,12 @@ class API_Storage_Ceph_Volume_Element(Resource):
             type: string
             required: false
             description: The new volume name
+          - in: query
+            name: force
+            type: boolean
+            required: false
+            default: flase
+            description: Force action if new volume size would violate 80% full soft cap on the pool
         responses:
           200:
             description: OK
@@ -5894,7 +5929,9 @@ class API_Storage_Ceph_Volume_Element(Resource):
             return {"message": "Can only perform one modification at once"}, 400
 
         if reqargs.get("new_size", None):
-            return api_helper.ceph_volume_resize(pool, volume, reqargs.get("new_size"))
+            return api_helper.ceph_volume_resize(
+                pool, volume, reqargs.get("new_size"), reqargs.get("force", False)
+            )
         if reqargs.get("new_name", None):
             return api_helper.ceph_volume_rename(pool, volume, reqargs.get("new_name"))
         return {"message": "At least one modification must be specified"}, 400
@@ -5935,7 +5972,11 @@ class API_Storage_Ceph_Volume_Element_Clone(Resource):
                 "name": "new_volume",
                 "required": True,
                 "helptext": "A new volume name must be specified.",
-            }
+            },
+            {
+                "name": "force",
+                "required": False,
+            },
         ]
     )
     @Authenticator
@@ -5951,6 +5992,12 @@ class API_Storage_Ceph_Volume_Element_Clone(Resource):
             type: string
             required: true
             description: The name of the new cloned volume
+          - in: query
+            name: force
+            type: boolean
+            required: false
+            default: flase
+            description: Force action if clone volume size would violate 80% full soft cap on the pool
         responses:
           200:
             description: OK
@@ -5969,7 +6016,7 @@ class API_Storage_Ceph_Volume_Element_Clone(Resource):
               id: Message
         """
         return api_helper.ceph_volume_clone(
-            pool, reqargs.get("new_volume", None), volume
+            pool, reqargs.get("new_volume", None), volume, reqargs.get("force", None)
         )
 
 

api-daemon/pvcapid/helper.py

@@ -1869,11 +1869,13 @@ def ceph_volume_list(zkhandler, pool=None, limit=None, is_fuzzy=True):
 
 
 @ZKConnection(config)
-def ceph_volume_add(zkhandler, pool, name, size):
+def ceph_volume_add(zkhandler, pool, name, size, force_flag):
     """
     Add a Ceph RBD volume to the PVC Ceph storage cluster.
     """
-    retflag, retdata = pvc_ceph.add_volume(zkhandler, pool, name, size)
+    retflag, retdata = pvc_ceph.add_volume(
+        zkhandler, pool, name, size, force_flag=force_flag
+    )
 
     if retflag:
         retcode = 200
@@ -1885,11 +1887,13 @@ def ceph_volume_add(zkhandler, pool, name, size):
 
 
 @ZKConnection(config)
-def ceph_volume_clone(zkhandler, pool, name, source_volume):
+def ceph_volume_clone(zkhandler, pool, name, source_volume, force_flag):
     """
     Clone a Ceph RBD volume to a new volume on the PVC Ceph storage cluster.
     """
-    retflag, retdata = pvc_ceph.clone_volume(zkhandler, pool, source_volume, name)
+    retflag, retdata = pvc_ceph.clone_volume(
+        zkhandler, pool, source_volume, name, force_flag=force_flag
+    )
 
     if retflag:
         retcode = 200
@@ -1901,11 +1905,13 @@ def ceph_volume_clone(zkhandler, pool, name, source_volume):
 
 
 @ZKConnection(config)
-def ceph_volume_resize(zkhandler, pool, name, size):
+def ceph_volume_resize(zkhandler, pool, name, size, force_flag):
     """
     Resize an existing Ceph RBD volume in the PVC Ceph storage cluster.
     """
-    retflag, retdata = pvc_ceph.resize_volume(zkhandler, pool, name, size)
+    retflag, retdata = pvc_ceph.resize_volume(
+        zkhandler, pool, name, size, force_flag=force_flag
+    )
 
     if retflag:
         retcode = 200

client-cli/pvc/cli/cli.py

@@ -687,7 +687,10 @@ def cli_cluster_task(task_id, wait_flag, format_function):
 
     if wait_flag:
         # First validate that this is actually a valid task that is running
+        echo(CLI_CONFIG, "Querying cluster for tasks...", newline=False)
         retcode, retdata = pvc.lib.common.task_status(CLI_CONFIG, None)
+        echo(CLI_CONFIG, " done.")
+        echo(CLI_CONFIG, "")
         if task_id in [i["id"] for i in retdata]:
             task = [i for i in retdata if i["id"] == task_id][0]
             retmsg = wait_for_celery_task(
@@ -699,7 +702,10 @@ def cli_cluster_task(task_id, wait_flag, format_function):
             retmsg = f"No task with ID {task_id} found."
         finish(retcode, retmsg)
     else:
+        echo(CLI_CONFIG, "Querying cluster for tasks...", newline=False)
         retcode, retdata = pvc.lib.common.task_status(CLI_CONFIG, task_id)
+        echo(CLI_CONFIG, " done.")
+        echo(CLI_CONFIG, "")
         finish(retcode, retdata, format_function)
 
 
@@ -4094,12 +4100,26 @@ def cli_storage_volume():
 @click.argument("pool")
 @click.argument("name")
 @click.argument("size")
-def cli_storage_volume_add(pool, name, size):
+@click.option(
+    "-f",
+    "--force",
+    "force_flag",
+    is_flag=True,
+    default=False,
+    help="Force creation even if volume would violate 80% full safe free space.",
+)
+def cli_storage_volume_add(pool, name, size, force_flag):
     """
     Add a new Ceph RBD volume in pool POOL with name NAME and size SIZE (in human units, e.g. 1024M, 20G, etc.).
+
+    PVC will prevent the creation of a volume who's size is greater than the available free space on the pool. This cannot be overridden.
+
+    PVC will prevent the creation of a volume who's size is greater than the 80% full safe free space on the pool. This can be overridden with the "-f"/"--force" option but this may be dangerous!
     """
 
-    retcode, retmsg = pvc.lib.storage.ceph_volume_add(CLI_CONFIG, pool, name, size)
+    retcode, retmsg = pvc.lib.storage.ceph_volume_add(
+        CLI_CONFIG, pool, name, size, force_flag=force_flag
+    )
     finish(retcode, retmsg)
 
 
@@ -4165,14 +4185,26 @@ def cli_storage_volume_remove(pool, name):
 @click.argument("pool")
 @click.argument("name")
 @click.argument("size")
+@click.option(
+    "-f",
+    "--force",
+    "force_flag",
+    is_flag=True,
+    default=False,
+    help="Force resize even if volume would violate 80% full safe free space.",
+)
 @confirm_opt("Resize volume {name} in pool {pool} to size {size}")
-def cli_storage_volume_resize(pool, name, size):
+def cli_storage_volume_resize(pool, name, size, force_flag):
     """
     Resize an existing Ceph RBD volume with name NAME in pool POOL to size SIZE (in human units, e.g. 1024M, 20G, etc.).
+
+    PVC will prevent the resize of a volume who's new size is greater than the available free space on the pool. This cannot be overridden.
+
+    PVC will prevent the resize of a volume who's new size is greater than the 80% full safe free space on the pool. This can be overridden with the "-f"/"--force" option but this may be dangerous!
     """
 
     retcode, retmsg = pvc.lib.storage.ceph_volume_modify(
-        CLI_CONFIG, pool, name, new_size=size
+        CLI_CONFIG, pool, name, new_size=size, force_flag=force_flag
     )
     finish(retcode, retmsg)
 
@@ -4205,13 +4237,25 @@ def cli_storage_volume_rename(pool, name, new_name):
 @click.argument("pool")
 @click.argument("name")
 @click.argument("new_name")
-def cli_storage_volume_clone(pool, name, new_name):
+@click.option(
+    "-f",
+    "--force",
+    "force_flag",
+    is_flag=True,
+    default=False,
+    help="Force clone even if volume would violate 80% full safe free space.",
+)
+def cli_storage_volume_clone(pool, name, new_name, force_flag):
     """
     Clone a Ceph RBD volume with name NAME in pool POOL to name NEW_NAME in pool POOL.
+
+    PVC will prevent the clone of a volume who's new size is greater than the available free space on the pool. This cannot be overridden.
+
+    PVC will prevent the clone of a volume who's new size is greater than the 80% full safe free space on the pool. This can be overridden with the "-f"/"--force" option but this may be dangerous!
     """
 
     retcode, retmsg = pvc.lib.storage.ceph_volume_clone(
-        CLI_CONFIG, pool, name, new_name
+        CLI_CONFIG, pool, name, new_name, force_flag=force_flag
     )
     finish(retcode, retmsg)
 

client-cli/pvc/cli/formatters.py

@@ -645,6 +645,24 @@ def cli_cluster_task_format_pretty(CLI_CONFIG, task_data):
         if _task_type_length > task_type_length:
             task_type_length = _task_type_length
 
+        for arg_name, arg_data in task["kwargs"].items():
+            # Skip the "run_on" argument
+            if arg_name == "run_on":
+                continue
+
+            # task_arg_name column
+            _task_arg_name_length = len(str(arg_name)) + 1
+            if _task_arg_name_length > task_arg_name_length:
+                task_arg_name_length = _task_arg_name_length
+
+    task_header_length = (
+        task_id_length + task_name_length + task_type_length + task_worker_length + 3
+    )
+    max_task_data_length = (
+        MAX_CONTENT_WIDTH - task_header_length - task_arg_name_length - 2
+    )
+
+    for task in task_data:
         updated_kwargs = list()
         for arg_name, arg_data in task["kwargs"].items():
             # Skip the "run_on" argument
@@ -656,15 +674,30 @@ def cli_cluster_task_format_pretty(CLI_CONFIG, task_data):
             if _task_arg_name_length > task_arg_name_length:
                 task_arg_name_length = _task_arg_name_length
 
-            if len(str(arg_data)) > 17:
-                arg_data = arg_data[:17] + "..."
+            if isinstance(arg_data, list):
+                for subarg_data in arg_data:
+                    if len(subarg_data) > max_task_data_length:
+                        subarg_data = (
+                            str(subarg_data[: max_task_data_length - 4]) + " ..."
+                        )
 
-            # task_arg_data column
-            _task_arg_data_length = len(str(arg_data)) + 1
-            if _task_arg_data_length > task_arg_data_length:
-                task_arg_data_length = _task_arg_data_length
+                    # task_arg_data column
+                    _task_arg_data_length = len(str(subarg_data)) + 1
+                    if _task_arg_data_length > task_arg_data_length:
+                        task_arg_data_length = _task_arg_data_length
+
+                    updated_kwargs.append({"name": arg_name, "data": subarg_data})
+            else:
+                if len(str(arg_data)) > 24:
+                    arg_data = str(arg_data[:24]) + " ..."
+
+                    # task_arg_data column
+                    _task_arg_data_length = len(str(arg_data)) + 1
+                    if _task_arg_data_length > task_arg_data_length:
+                        task_arg_data_length = _task_arg_data_length
+
+                updated_kwargs.append({"name": arg_name, "data": arg_data})
 
-            updated_kwargs.append({"name": arg_name, "data": arg_data})
         task["kwargs"] = updated_kwargs
         tasks.append(task)
 

client-cli/pvc/cli/helpers.py

@@ -246,6 +246,8 @@ def vm_autobackup(
     Perform automatic backups of VMs based on an external config file.
     """
 
+    backup_summary = dict()
+
     if email_report is not None:
         from email.utils import formatdate
         from socket import gethostname
@@ -553,6 +555,8 @@ def vm_autobackup(
         with open(autobackup_state_file, "w") as fh:
             jdump(state_data, fh)
 
+        backup_summary[vm] = tracked_backups
+
     if autobackup_config["auto_mount_enabled"]:
         # Execute each unmount_cmds command in sequence
         for cmd in autobackup_config["unmount_cmds"]:
@@ -588,20 +592,6 @@ def vm_autobackup(
     if email_report is not None:
         echo(CLI_CONFIG, "")
         echo(CLI_CONFIG, f"Sending email summary report to {email_report}")
-        backup_summary = dict()
-        for vm in backup_vms:
-            backup_path = f"{backup_suffixed_path}/{vm}"
-            autobackup_state_file = f"{backup_path}/.autobackup.json"
-            if not path.exists(backup_path) or not path.exists(autobackup_state_file):
-                # There are no new backups so the list is empty
-                state_data = dict()
-                tracked_backups = list()
-            else:
-                with open(autobackup_state_file) as fh:
-                    state_data = jload(fh)
-                tracked_backups = state_data["tracked_backups"]
-
-            backup_summary[vm] = tracked_backups
 
         current_datetime = datetime.now()
         email_datetime = formatdate(float(current_datetime.strftime("%s")))

client-cli/pvc/cli/waiters.py

@@ -115,6 +115,8 @@ def wait_for_celery_task(CLI_CONFIG, task_detail, start_late=False):
         )
         while True:
             sleep(0.5)
+            if isinstance(task_status, tuple):
+                continue
             if task_status.get("state") != "RUNNING":
                 break
             if task_status.get("current") > last_task:

client-cli/pvc/lib/common.py

@@ -140,15 +140,31 @@ def call_api(
     # Determine the request type and hit the API
     disable_warnings()
     try:
+        response = None
         if operation == "get":
-            response = requests.get(
-                uri,
-                timeout=timeout,
-                headers=headers,
-                params=params,
-                data=data,
-                verify=config["verify_ssl"],
-            )
+            retry_on_code = [429, 500, 502, 503, 504]
+            for i in range(3):
+                failed = False
+                try:
+                    response = requests.get(
+                        uri,
+                        timeout=timeout,
+                        headers=headers,
+                        params=params,
+                        data=data,
+                        verify=config["verify_ssl"],
+                    )
+                    if response.status_code in retry_on_code:
+                        failed = True
+                        continue
+                except requests.exceptions.ConnectionError:
+                    failed = True
+                    pass
+            if failed:
+                error = f"Code {response.status_code}" if response else "Timeout"
+                raise requests.exceptions.ConnectionError(
+                    f"Failed to connect after 3 tries ({error})"
+                )
         if operation == "post":
             response = requests.post(
                 uri,
@@ -189,7 +205,8 @@ def call_api(
             )
     except Exception as e:
         message = "Failed to connect to the API: {}".format(e)
-        response = ErrorResponse({"message": message}, 500)
+        code = response.status_code if response else 504
+        response = ErrorResponse({"message": message}, code)
 
     # Display debug output
     if config["debug"]:

client-cli/pvc/lib/storage.py

@@ -430,7 +430,7 @@ def format_list_osd(config, osd_list):
             )
             continue
 
-        if osd_information["is_split"]:
+        if osd_information.get("is_split") is not None:
             osd_information["device"] = f"{osd_information['device']} [s]"
 
         # Deal with the size to human readable
@@ -1172,15 +1172,15 @@ def ceph_volume_list(config, limit, pool):
         return False, response.json().get("message", "")
 
 
-def ceph_volume_add(config, pool, volume, size):
+def ceph_volume_add(config, pool, volume, size, force_flag=False):
     """
     Add new Ceph volume
 
     API endpoint: POST /api/v1/storage/ceph/volume
-    API arguments: volume={volume}, pool={pool}, size={size}
+    API arguments: volume={volume}, pool={pool}, size={size}, force={force_flag}
     API schema: {"message":"{data}"}
     """
-    params = {"volume": volume, "pool": pool, "size": size}
+    params = {"volume": volume, "pool": pool, "size": size, "force": force_flag}
     response = call_api(config, "post", "/storage/ceph/volume", params=params)
 
     if response.status_code == 200:
@@ -1261,12 +1261,14 @@ def ceph_volume_remove(config, pool, volume):
     return retstatus, response.json().get("message", "")
 
 
-def ceph_volume_modify(config, pool, volume, new_name=None, new_size=None):
+def ceph_volume_modify(
+    config, pool, volume, new_name=None, new_size=None, force_flag=False
+):
     """
     Modify Ceph volume
 
     API endpoint: PUT /api/v1/storage/ceph/volume/{pool}/{volume}
-    API arguments:
+    API arguments: [new_name={new_name}], [new_size={new_size}], force_flag={force_flag}
     API schema: {"message":"{data}"}
     """
 
@@ -1275,6 +1277,7 @@ def ceph_volume_modify(config, pool, volume, new_name=None, new_size=None):
         params["new_name"] = new_name
     if new_size:
         params["new_size"] = new_size
+        params["force"] = force_flag
 
     response = call_api(
         config,
@@ -1291,15 +1294,15 @@ def ceph_volume_modify(config, pool, volume, new_name=None, new_size=None):
     return retstatus, response.json().get("message", "")
 
 
-def ceph_volume_clone(config, pool, volume, new_volume):
+def ceph_volume_clone(config, pool, volume, new_volume, force_flag=False):
     """
     Clone Ceph volume
 
     API endpoint: POST /api/v1/storage/ceph/volume/{pool}/{volume}
-    API arguments: new_volume={new_volume
+    API arguments: new_volume={new_volume, force_flag={force_flag}
     API schema: {"message":"{data}"}
     """
-    params = {"new_volume": new_volume}
+    params = {"new_volume": new_volume, "force_flag": force_flag}
     response = call_api(
         config,
         "post",

client-cli/pvc/lib/vm.py

@@ -89,6 +89,7 @@ def vm_define(
     node_selector,
     node_autostart,
     migration_method,
+    migration_max_downtime,
     user_tags,
     protected_tags,
 ):
@@ -96,7 +97,7 @@ def vm_define(
     Define a new VM on the cluster
 
     API endpoint: POST /vm
-    API arguments: xml={xml}, node={node}, limit={node_limit}, selector={node_selector}, autostart={node_autostart}, migration_method={migration_method}, user_tags={user_tags}, protected_tags={protected_tags}
+    API arguments: xml={xml}, node={node}, limit={node_limit}, selector={node_selector}, autostart={node_autostart}, migration_method={migration_method}, migration_max_downtime={migration_max_downtime}, user_tags={user_tags}, protected_tags={protected_tags}
     API schema: {"message":"{data}"}
     """
     params = {
@@ -105,6 +106,7 @@ def vm_define(
         "selector": node_selector,
         "autostart": node_autostart,
         "migration_method": migration_method,
+        "migration_max_downtime": migration_max_downtime,
         "user_tags": user_tags,
         "protected_tags": protected_tags,
     }
@@ -1630,6 +1632,7 @@ def format_info(config, domain_information, long_output):
         "migrate": ansiprint.blue(),
         "unmigrate": ansiprint.blue(),
         "provision": ansiprint.blue(),
+        "restore": ansiprint.blue(),
     }
     ainformation.append(
         "{}State:{}              {}{}{}".format(
@@ -1714,7 +1717,7 @@ def format_info(config, domain_information, long_output):
         "{}Max live downtime:{}  {}".format(
             ansiprint.purple(),
             ansiprint.end(),
-            f"{domain_information['migration_max_downtime']} ms",
+            f"{domain_information.get('migration_max_downtime')} ms",
         )
     )
 

client-cli/setup.py

@@ -2,7 +2,7 @@ from setuptools import setup
 
 setup(
     name="pvc",
-    version="0.9.90",
+    version="0.9.94",
     packages=["pvc.cli", "pvc.lib"],
     install_requires=[
         "Click",

daemon-common/ceph.py

@@ -553,7 +553,7 @@ def getVolumeInformation(zkhandler, pool, volume):
     return volume_information
 
 
-def add_volume(zkhandler, pool, name, size):
+def add_volume(zkhandler, pool, name, size, force_flag=False):
     # 1. Verify the size of the volume
     pool_information = getPoolInformation(zkhandler, pool)
     size_bytes = format_bytes_fromhuman(size)
@@ -563,12 +563,27 @@ def add_volume(zkhandler, pool, name, size):
             f"ERROR: Requested volume size '{size}' does not have a valid SI unit",
         )
 
-    if size_bytes >= int(pool_information["stats"]["free_bytes"]):
+    pool_total_free_bytes = int(pool_information["stats"]["free_bytes"])
+    if size_bytes >= pool_total_free_bytes:
         return (
             False,
             f"ERROR: Requested volume size '{format_bytes_tohuman(size_bytes)}' is greater than the available free space in the pool ('{format_bytes_tohuman(pool_information['stats']['free_bytes'])}')",
         )
 
+    # Check if we're greater than 80% utilization after the create; error if so unless we have the force flag
+    pool_total_bytes = (
+        int(pool_information["stats"]["used_bytes"]) + pool_total_free_bytes
+    )
+    pool_safe_total_bytes = int(pool_total_bytes * 0.80)
+    pool_safe_free_bytes = pool_safe_total_bytes - int(
+        pool_information["stats"]["used_bytes"]
+    )
+    if size_bytes >= pool_safe_free_bytes and not force_flag:
+        return (
+            False,
+            f"ERROR: Requested volume size '{format_bytes_tohuman(size_bytes)}' is greater than the safe free space in the pool ('{format_bytes_tohuman(pool_safe_free_bytes)}' for 80% full); retry with force to ignore this error",
+        )
+
     # 2. Create the volume
     retcode, stdout, stderr = common.run_os_command(
         "rbd create --size {}B {}/{}".format(size_bytes, pool, name)
@@ -596,13 +611,39 @@ def add_volume(zkhandler, pool, name, size):
     )
 
 
-def clone_volume(zkhandler, pool, name_src, name_new):
+def clone_volume(zkhandler, pool, name_src, name_new, force_flag=False):
+    # 1. Verify the volume
     if not verifyVolume(zkhandler, pool, name_src):
         return False, 'ERROR: No volume with name "{}" is present in pool "{}".'.format(
             name_src, pool
         )
 
-    # 1. Clone the volume
+    volume_stats_raw = zkhandler.read(("volume.stats", f"{pool}/{name_src}"))
+    volume_stats = dict(json.loads(volume_stats_raw))
+    size_bytes = volume_stats["size"]
+    pool_information = getPoolInformation(zkhandler, pool)
+    pool_total_free_bytes = int(pool_information["stats"]["free_bytes"])
+    if size_bytes >= pool_total_free_bytes:
+        return (
+            False,
+            f"ERROR: Clone volume size '{format_bytes_tohuman(size_bytes)}' is greater than the available free space in the pool ('{format_bytes_tohuman(pool_information['stats']['free_bytes'])}')",
+        )
+
+    # Check if we're greater than 80% utilization after the create; error if so unless we have the force flag
+    pool_total_bytes = (
+        int(pool_information["stats"]["used_bytes"]) + pool_total_free_bytes
+    )
+    pool_safe_total_bytes = int(pool_total_bytes * 0.80)
+    pool_safe_free_bytes = pool_safe_total_bytes - int(
+        pool_information["stats"]["used_bytes"]
+    )
+    if size_bytes >= pool_safe_free_bytes and not force_flag:
+        return (
+            False,
+            f"ERROR: Clone volume size '{format_bytes_tohuman(size_bytes)}' is greater than the safe free space in the pool ('{format_bytes_tohuman(pool_safe_free_bytes)}' for 80% full); retry with force to ignore this error",
+        )
+
+    # 2. Clone the volume
     retcode, stdout, stderr = common.run_os_command(
         "rbd copy {}/{} {}/{}".format(pool, name_src, pool, name_new)
     )
@@ -614,13 +655,13 @@ def clone_volume(zkhandler, pool, name_src, name_new):
             ),
         )
 
-    # 2. Get volume stats
+    # 3. Get volume stats
     retcode, stdout, stderr = common.run_os_command(
         "rbd info --format json {}/{}".format(pool, name_new)
     )
     volstats = stdout
 
-    # 3. Add the new volume to Zookeeper
+    # 4. Add the new volume to Zookeeper
     zkhandler.write(
         [
             (("volume", f"{pool}/{name_new}"), ""),
@@ -634,7 +675,7 @@ def clone_volume(zkhandler, pool, name_src, name_new):
     )
 
 
-def resize_volume(zkhandler, pool, name, size):
+def resize_volume(zkhandler, pool, name, size, force_flag=False):
     if not verifyVolume(zkhandler, pool, name):
         return False, 'ERROR: No volume with name "{}" is present in pool "{}".'.format(
             name, pool
@@ -649,12 +690,27 @@ def resize_volume(zkhandler, pool, name, size):
             f"ERROR: Requested volume size '{size}' does not have a valid SI unit",
         )
 
-    if size_bytes >= int(pool_information["stats"]["free_bytes"]):
+    pool_total_free_bytes = int(pool_information["stats"]["free_bytes"])
+    if size_bytes >= pool_total_free_bytes:
         return (
             False,
             f"ERROR: Requested volume size '{format_bytes_tohuman(size_bytes)}' is greater than the available free space in the pool ('{format_bytes_tohuman(pool_information['stats']['free_bytes'])}')",
         )
 
+    # Check if we're greater than 80% utilization after the create; error if so unless we have the force flag
+    pool_total_bytes = (
+        int(pool_information["stats"]["used_bytes"]) + pool_total_free_bytes
+    )
+    pool_safe_total_bytes = int(pool_total_bytes * 0.80)
+    pool_safe_free_bytes = pool_safe_total_bytes - int(
+        pool_information["stats"]["used_bytes"]
+    )
+    if size_bytes >= pool_safe_free_bytes and not force_flag:
+        return (
+            False,
+            f"ERROR: Requested volume size '{format_bytes_tohuman(size_bytes)}' is greater than the safe free space in the pool ('{format_bytes_tohuman(pool_safe_free_bytes)}' for 80% full); retry with force to ignore this error",
+        )
+
     # 2. Resize the volume
     retcode, stdout, stderr = common.run_os_command(
         "rbd resize --size {} {}/{}".format(

daemon-common/cluster.py

@@ -1201,7 +1201,7 @@ def get_resource_metrics(zkhandler):
         try:
             user_time = vm["vcpu_stats"]["user_time"] / 1000000
         except Exception:
-            cpu_time = 0
+            user_time = 0
         output_lines.append(
             f"pvc_vm_vcpus_user_time{{vm=\"{vm['name']}\"}} {user_time}"
         )
@@ -1230,7 +1230,7 @@ def get_resource_metrics(zkhandler):
     )
     output_lines.append("# TYPE pvc_vm_memory_stats_actual gauge")
     for vm in vm_data:
-        actual_memory = vm["memory_stats"]["actual"]
+        actual_memory = vm["memory_stats"].get("actual", 0)
         output_lines.append(
             f"pvc_vm_memory_stats_actual{{vm=\"{vm['name']}\"}} {actual_memory}"
         )
@@ -1238,7 +1238,7 @@ def get_resource_metrics(zkhandler):
     output_lines.append("# HELP pvc_vm_memory_stats_rss PVC VM RSS memory KB")
     output_lines.append("# TYPE pvc_vm_memory_stats_rss gauge")
     for vm in vm_data:
-        rss_memory = vm["memory_stats"]["rss"]
+        rss_memory = vm["memory_stats"].get("rss", 0)
         output_lines.append(
             f"pvc_vm_memory_stats_rss{{vm=\"{vm['name']}\"}} {rss_memory}"
         )

daemon-common/zkhandler.py

@@ -57,10 +57,11 @@ class ZKConnection(object):
                 schema_version = 0
             zkhandler.schema.load(schema_version, quiet=True)
 
-            ret = function(zkhandler, *args, **kwargs)
-
-            zkhandler.disconnect()
-            del zkhandler
+            try:
+                ret = function(zkhandler, *args, **kwargs)
+            finally:
+                zkhandler.disconnect()
+                del zkhandler
 
             return ret
 

debian/changelog

@@ -1,3 +1,38 @@
+pvc (0.9.94-0) unstable; urgency=high
+
+  * [CLI Client] Fixes an incorrect ordering issue with autobackup summary emails
+  * [API Daemon/CLI Client] Adds an additional safety check for 80% cluster fullness when doing volume adds or resizes
+  * [API Daemon/CLI Client] Adds safety checks to volume clones as well
+  * [API Daemon] Fixes a few remaining memory bugs for stopped/disabled VMs
+
+ -- Joshua M. Boniface <joshua@boniface.me>  Mon, 05 Feb 2024 09:58:07 -0500
+
+pvc (0.9.93-0) unstable; urgency=high
+
+  * [API Daemon] Fixes a bug where stuck zkhandler threads were not cleaned up on error
+
+ -- Joshua M. Boniface <joshua@boniface.me>  Tue, 30 Jan 2024 09:51:21 -0500
+
+pvc (0.9.92-0) unstable; urgency=high
+
+  * [CLI Client] Adds the new restore state to the colours list for VM status
+  * [API Daemon] Fixes an incorrect variable assignment
+  * [Provisioner] Improves the error handling of various steps in the debootstrap and rinse example scripts
+  * [CLI Client] Fixes two bugs around missing keys that were added recently (uses get() instead direct dictionary refs)
+  * [CLI Client] Improves API error handling via GET retries (x3) and better server status code handling
+
+ -- Joshua M. Boniface <joshua@boniface.me>  Mon, 29 Jan 2024 09:39:10 -0500
+
+pvc (0.9.91-0) unstable; urgency=high
+
+  * [Client CLI] Fixes a bug and improves output during cluster task events.
+  * [Client CLI] Improves the output of the task list display.
+  * [Provisioner] Fixes some missing cloud-init modules in the default debootstrap script.
+  * [Client CLI] Fixes a bug with a missing argument to the vm_define helper function.
+  * [All] Fixes inconsistent package find + rm commands to avoid errors in dpkg.
+
+ -- Joshua M. Boniface <joshua@boniface.me>  Tue, 23 Jan 2024 10:02:19 -0500
+
 pvc (0.9.90-0) unstable; urgency=high
 
   * [Client CLI/API Daemon] Adds additional backup metainfo and an emailed report option to autobackups.

debian/pvc-client-cli.postinst

@@ -2,7 +2,12 @@
 
 # Generate the bash completion configuration
 if [ -d /etc/bash_completion.d ]; then
+    echo "Installing BASH completion configuration"
     _PVC_COMPLETE=source_bash pvc > /etc/bash_completion.d/pvc
 fi
 
+# Remove any cached CPython directories or files
+echo "Cleaning up CPython caches"
+find /usr/lib/python3/dist-packages/pvc -type d -name "__pycache__" -exec rm -fr {} + &>/dev/null || true
+
 exit 0

debian/pvc-daemon-api.preinst

@@ -1,5 +1,5 @@
 #!/bin/sh
 
 # Remove any cached CPython directories or files
-echo "Cleaning up existing CPython files"
-find /usr/share/pvc/pvcapid -type d -name "__pycache__" -exec rm -rf {} \; &>/dev/null || true
+echo "Cleaning up CPython caches"
+find /usr/share/pvc/pvcapid -type d -name "__pycache__" -exec rm -fr {} + &>/dev/null || true

debian/pvc-daemon-common.preinst

@@ -0,0 +1,5 @@
+#!/bin/sh
+
+# Remove any cached CPython directories or files
+echo "Cleaning up CPython caches"
+find /usr/share/pvc/daemon_lib -type d -name "__pycache__" -exec rm -fr {} + &>/dev/null || true

debian/pvc-daemon-health.preinst

@@ -1,6 +1,6 @@
 #!/bin/sh
 
 # Remove any cached CPython directories or files
-echo "Cleaning up existing CPython files"
-find /usr/share/pvc/pvchealthd -type d -name "__pycache__" -exec rm -rf {} \; &>/dev/null || true
-find /usr/share/pvc/plugins -type d -name "__pycache__" -exec rm -rf {} \; &>/dev/null || true
+echo "Cleaning up CPython caches"
+find /usr/share/pvc/pvchealthd -type d -name "__pycache__" -exec rm -fr {} + &>/dev/null || true
+find /usr/share/pvc/plugins -type d -name "__pycache__" -exec rm -fr {} + &>/dev/null || true

debian/pvc-daemon-node.preinst

@@ -1,5 +1,5 @@
 #!/bin/sh
 
 # Remove any cached CPython directories or files
-echo "Cleaning up existing CPython files"
-find /usr/share/pvc/pvcnoded -type d -name "__pycache__" -exec rm -rf {} \; &>/dev/null || true
+echo "Cleaning up CPython caches"
+find /usr/share/pvc/pvcnoded -type d -name "__pycache__" -exec rm -fr {} + &>/dev/null || true

debian/pvc-daemon-worker.preinst

@@ -1,5 +1,5 @@
 #!/bin/sh
 
 # Remove any cached CPython directories or files
-echo "Cleaning up existing CPython files"
-find /usr/share/pvc/pvcworkerd -type d -name "__pycache__" -exec rm -rf {} \; &>/dev/null || true
+echo "Cleaning up CPython caches"
+find /usr/share/pvc/pvcworkerd -type d -name "__pycache__" -exec rm -fr {} + &>/dev/null || true

debian/rules

@@ -13,7 +13,7 @@ override_dh_python3:
 	rm -r $(CURDIR)/client-cli/.pybuild $(CURDIR)/client-cli/pvc.egg-info
 
 override_dh_auto_clean:
-	find . -name "__pycache__" -o -name ".pybuild" -exec rm -r {} \; || true
+	find . -name "__pycache__" -o -name ".pybuild" -exec rm -fr {} + || true
 
 # If you need to rebuild the Sphinx documentation
 # Add spinxdoc to the dh --with line

health-daemon/pvchealthd/Daemon.py

@@ -33,7 +33,7 @@ import os
 import signal
 
 # Daemon version
-version = "0.9.90"
+version = "0.9.94"
 
 
 ##########################################################

node-daemon/pvcnoded/Daemon.py

@@ -49,7 +49,7 @@ import re
 import json
 
 # Daemon version
-version = "0.9.90"
+version = "0.9.94"
 
 
 ##########################################################

worker-daemon/pvcworkerd/Daemon.py

@@ -44,7 +44,7 @@ from daemon_lib.vmbuilder import (
 )
 
 # Daemon version
-version = "0.9.90"
+version = "0.9.94"
 
 
 config = cfg.get_configuration()