[vm-repair] Adding repair-and-restore one command flow for fstab scripts (#6244)

Author: haagha

src/vm-repair/HISTORY.rst

@@ -2,6 +2,10 @@
 Release History
 ===============
 
+0.5.4
+++++++
+Adding repair-and-restore command to create a one command flow for vm-repair with fstab scripts.
+
 0.5.3
 ++++++
 Removing check for EncryptionSettingsCollection.enabled is string 'false'.
@@ -18,6 +22,10 @@ Updated exsiting privateIpAddress field to privateIPAddress and privateIpAllocat
 ++++++
 Support for hosting repair vm in existing resource group and fixing existing resource group logic 
 
+0.5.0
+++++++
+Support for hosting repair vm in existing resource group and fixing existing resource group logic 
+
 0.4.10
 ++++++
 Support for hosting repair vm in existing resource group and fixing existing resource group logic 

src/vm-repair/azext_vm_repair/_help.py

@@ -92,3 +92,12 @@
           text: >
             az vm repair reset-nic -g MyResourceGroup -n MyVM --yes --verbose
 """
+
+helps['vm repair repair-and-restore'] = """
+    type: command
+    short-summary: Repair and restore the VM.
+    examples:
+        - name: Repair and restore a VM.
+          text: >
+            az vm repair repair-and-restore --name vmrepairtest --resource-group MyResourceGroup --verbose
+"""

src/vm-repair/azext_vm_repair/_params.py

@@ -53,3 +53,11 @@ def load_arguments(self, _):
     with self.argument_context('vm repair reset-nic') as c:
         c.argument('subscriptionid', help='Subscription id to default subscription using `az account set -s NAME_OR_ID`.')
         c.argument('yes', help='Do not prompt for confirmation to start VM if it is not running.')
+
+    with self.argument_context('vm repair repair-and-restore') as c:
+        c.argument('repair_username', help='Admin username for repair VM.')
+        c.argument('repair_password', help='Admin password for the repair VM.')
+        c.argument('copy_disk_name', help='Name of OS disk copy.')
+        c.argument('repair_vm_name', help='Name of repair VM.')
+        c.argument('copy_disk_name', help='Name of OS disk copy.')
+        c.argument('repair_group_name', help='Name for new or existing resource group that will contain repair VM.')

src/vm-repair/azext_vm_repair/_validators.py

@@ -376,3 +376,57 @@ def validate_vm_username(username, is_linux):
 
     if username.lower() in disallowed_user_names:
         raise CLIError("This username '{}' meets the general requirements, but is specifically disallowed. Please try a different value.".format(username))
+
+
+def validate_repair_and_restore(cmd, namespace):
+    check_extension_version(EXTENSION_NAME)
+
+    logger.info('Validating repair and restore parameters...')
+
+    logger.info(namespace.vm_name + ' ' + namespace.resource_group_name)
+
+    # Check if VM exists and is not classic VM
+    source_vm = _validate_and_get_vm(cmd, namespace.resource_group_name, namespace.vm_name)
+    is_linux = _is_linux_os(source_vm)
+
+    # Check repair vm name
+    namespace.repair_vm_name = ('repair-' + namespace.vm_name)[:14] + '_'
+    logger.info('Repair VM name: %s', namespace.repair_vm_name)
+
+    # Check copy disk name
+    timestamp = datetime.utcnow().strftime('%Y%m%d%H%M%S')
+    if namespace.copy_disk_name:
+        _validate_disk_name(namespace.copy_disk_name)
+    else:
+        namespace.copy_disk_name = namespace.vm_name + '-DiskCopy-' + timestamp
+        logger.info('Copy disk name: %s', namespace.copy_disk_name)
+
+    # Check copy resouce group name
+    if namespace.repair_group_name:
+        if namespace.repair_group_name == namespace.resource_group_name:
+            raise CLIError('The repair resource group name cannot be the same as the source VM resource group.')
+        _validate_resource_group_name(namespace.repair_group_name)
+    else:
+        namespace.repair_group_name = 'repair-' + namespace.vm_name + '-' + timestamp
+        logger.info('Repair resource group name: %s', namespace.repair_group_name)
+
+    # Check encrypted disk
+    encryption_type, _, _, _ = _fetch_encryption_settings(source_vm)
+    # Currently only supporting single pass
+    if encryption_type in (Encryption.SINGLE_WITH_KEK, Encryption.SINGLE_WITHOUT_KEK):
+        if not namespace.unlock_encrypted_vm:
+            _prompt_encrypted_vm(namespace)
+    elif encryption_type is Encryption.DUAL:
+        logger.warning('The source VM\'s OS disk is encrypted using dual pass method.')
+        raise CLIError('The current command does not support VMs which were encrypted using dual pass.')
+    else:
+        logger.debug('The source VM\'s OS disk is not encrypted')
+
+    validate_vm_username(namespace.repair_username, is_linux)
+    validate_vm_password(namespace.repair_password, is_linux)
+    # Prompt input for public ip usage
+    namespace.associate_public_ip = False
+
+    # Validate repair run command
+    source_vm = _validate_and_get_vm(cmd, namespace.resource_group_name, namespace.vm_name)
+    is_linux = _is_linux_os(source_vm)

src/vm-repair/azext_vm_repair/commands.py

@@ -16,3 +16,4 @@ def load_command_table(self, _):
         g.custom_command('run', 'run', validator=validate_run)
         g.custom_command('list-scripts', 'list_scripts')
         g.custom_command('reset-nic', 'reset_nic', is_preview=True, validator=validate_reset_nic)
+        g.custom_command('repair-and-restore', 'repair_and_restore', is_preview=True)

src/vm-repair/azext_vm_repair/custom.py

@@ -14,7 +14,7 @@
 from azure.cli.command_modules.vm.custom import get_vm, _is_linux_os
 from azure.cli.command_modules.storage.storage_url_helpers import StorageResourceIdentifier
 from msrestazure.tools import parse_resource_id
-from .exceptions import SkuDoesNotSupportHyperV
+from .exceptions import AzCommandError, SkuNotAvailableError, UnmanagedDiskCopyError, WindowsOsNotAvailableError, RunScriptNotFoundForIdError, SkuDoesNotSupportHyperV, ScriptReturnsError, SupportingResourceNotFoundError, CommandCanceledByUserError
 
 from .command_helper_class import command_helper
 from .repair_utils import (
@@ -45,11 +45,15 @@
     _check_n_start_vm,
     _check_existing_rg
 )
-from .exceptions import AzCommandError, SkuNotAvailableError, UnmanagedDiskCopyError, WindowsOsNotAvailableError, RunScriptNotFoundForIdError, SkuDoesNotSupportHyperV, ScriptReturnsError, SupportingResourceNotFoundError, CommandCanceledByUserError
+from .exceptions import AzCommandError, RunScriptNotFoundForIdError, SupportingResourceNotFoundError, CommandCanceledByUserError
 logger = get_logger(__name__)
 
 
 def create(cmd, vm_name, resource_group_name, repair_password=None, repair_username=None, repair_vm_name=None, copy_disk_name=None, repair_group_name=None, unlock_encrypted_vm=False, enable_nested=False, associate_public_ip=False, distro='ubuntu', yes=False):
+
+    # log all the parameters
+    logger.debug('vm repair create command parameters: vm_name: %s, resource_group_name: %s, repair_password: %s, repair_username: %s, repair_vm_name: %s, copy_disk_name: %s, repair_group_name: %s, unlock_encrypted_vm: %s, enable_nested: %s, associate_public_ip: %s, distro: %s, yes: %s', vm_name, resource_group_name, repair_password, repair_username, repair_vm_name, copy_disk_name, repair_group_name, unlock_encrypted_vm, enable_nested, associate_public_ip, distro, yes)
+
     # Init command helper object
     command = command_helper(logger, cmd, 'vm repair create')
     # Main command calling block
@@ -101,7 +105,8 @@ def create(cmd, vm_name, resource_group_name, repair_password=None, repair_usern
             create_repair_vm_command += ' --zone {zone}'.format(zone=zone)
 
         # Create new resource group
-        if not _check_existing_rg(repair_group_name):
+        existing_rg = _check_existing_rg(repair_group_name)
+        if not existing_rg:
             create_resource_group_command = 'az group create -l {loc} -n {group_name}' \
                                             .format(loc=source_vm.location, group_name=repair_group_name)
             logger.info('Creating resource group for repair VM and its resources...')
@@ -272,7 +277,7 @@ def create(cmd, vm_name, resource_group_name, repair_password=None, repair_usern
     if not command.is_status_success():
         command.set_status_error()
         return_dict = command.init_return_dict()
-        if _check_existing_rg(repair_group_name):
+        if existing_rg:
             _clean_up_resources(repair_group_name, confirm=True)
         else:
             _clean_up_resources(repair_group_name, confirm=False)
@@ -305,9 +310,11 @@ def restore(cmd, vm_name, resource_group_name, disk_name=None, repair_vm_id=None
         # Fetch source and repair VM data
         source_vm = get_vm(cmd, resource_group_name, vm_name)
         is_managed = _uses_managed_disk(source_vm)
-        repair_vm_id = parse_resource_id(repair_vm_id)
-        repair_vm_name = repair_vm_id['name']
-        repair_resource_group = repair_vm_id['resource_group']
+        if repair_vm_id:
+            logger.info('Repair VM ID: %s', repair_vm_id)
+            repair_vm_id = parse_resource_id(repair_vm_id)
+            repair_vm_name = repair_vm_id['name']
+            repair_resource_group = repair_vm_id['resource_group']
         source_disk = None
 
         # MANAGED DISK
@@ -379,6 +386,10 @@ def restore(cmd, vm_name, resource_group_name, disk_name=None, repair_vm_id=None
 
 def run(cmd, vm_name, resource_group_name, run_id=None, repair_vm_id=None, custom_script_file=None, parameters=None, run_on_repair=False, preview=None):
 
+    # log method parameters
+    logger.debug('vm repair run parameters: vm_name: %s, resource_group_name: %s, run_id: %s, repair_vm_id: %s, custom_script_file: %s, parameters: %s, run_on_repair: %s, preview: %s',
+                 vm_name, resource_group_name, run_id, repair_vm_id, custom_script_file, parameters, run_on_repair, preview)
+
     # Init command helper object
     command = command_helper(logger, cmd, 'vm repair run')
     LINUX_RUN_SCRIPT_NAME = 'linux-run-driver.sh'
@@ -397,9 +408,13 @@ def run(cmd, vm_name, resource_group_name, run_id=None, repair_vm_id=None, custo
             script_name = WINDOWS_RUN_SCRIPT_NAME
 
         # If run_on_repair is False, then repair_vm is the source_vm (scripts run directly on source vm)
-        repair_vm_id = parse_resource_id(repair_vm_id)
-        repair_vm_name = repair_vm_id['name']
-        repair_resource_group = repair_vm_id['resource_group']
+        if run_on_repair:
+            repair_vm_id = parse_resource_id(repair_vm_id)
+            repair_vm_name = repair_vm_id['name']
+            repair_resource_group = repair_vm_id['resource_group']
+        else:
+            repair_vm_name = vm_name
+            repair_resource_group = resource_group_name
 
         run_command_params = []
         additional_scripts = []
@@ -650,3 +665,70 @@ def reset_nic(cmd, vm_name, resource_group_name, yes=False):
         return_dict = command.init_return_dict()
 
     return return_dict
+
+
+def repair_and_restore(cmd, vm_name, resource_group_name, repair_password=None, repair_username=None, repair_vm_name=None, copy_disk_name=None, repair_group_name=None):
+    from datetime import datetime
+    import secrets
+    import string
+
+    # Init command helper object
+    command = command_helper(logger, cmd, 'vm repair repair-and-restore')
+
+    password_length = 30
+    password_characters = string.ascii_lowercase + string.digits + string.ascii_uppercase
+    repair_password = ''.join(secrets.choice(password_characters) for i in range(password_length))
+
+    username_length = 20
+    username_characters = string.ascii_lowercase + string.digits
+    repair_username = ''.join(secrets.choice(username_characters) for i in range(username_length))
+
+    timestamp = datetime.utcnow().strftime('%Y%m%d%H%M%S')
+    repair_vm_name = ('repair-' + vm_name)[:14] + '_'
+    copy_disk_name = vm_name + '-DiskCopy-' + timestamp
+    repair_group_name = 'repair-' + vm_name + '-' + timestamp
+    existing_rg = _check_existing_rg(repair_group_name)
+
+    create_out = create(cmd, vm_name, resource_group_name, repair_password, repair_username, repair_vm_name=repair_vm_name, copy_disk_name=copy_disk_name, repair_group_name=repair_group_name, associate_public_ip=False, yes=True)
+
+    # log create_out
+    logger.info('create_out: %s', create_out)
+
+    repair_vm_name = create_out['repair_vm_name']
+    copy_disk_name = create_out['copied_disk_name']
+    repair_group_name = create_out['repair_resource_group']
+
+    logger.info('Running fstab run command')
+
+    try:
+        run_out = run(cmd, repair_vm_name, repair_group_name, run_id='linux-alar2', parameters=["fstab"])
+
+    except Exception:
+        command.set_status_error()
+        command.error_stack_trace = traceback.format_exc()
+        command.error_message = "Command failed when running fstab script."
+        command.message = "Command failed when running fstab script."
+        if existing_rg:
+            _clean_up_resources(repair_group_name, confirm=True)
+        else:
+            _clean_up_resources(repair_group_name, confirm=False)
+        return
+
+    # log run_out
+    logger.info('run_out: %s', run_out)
+
+    if run_out['script_status'] == 'ERROR':
+        logger.error('fstab script returned an error.')
+        if existing_rg:
+            _clean_up_resources(repair_group_name, confirm=True)
+        else:
+            _clean_up_resources(repair_group_name, confirm=False)
+        return
+
+    logger.info('Running restore command')
+    show_vm_id = 'az vm show -g {g} -n {n} --query id -o tsv' \
+        .format(g=repair_group_name, n=repair_vm_name)
+
+    repair_vm_id = _call_az_command(show_vm_id)
+
+    restore(cmd, vm_name, resource_group_name, copy_disk_name, repair_vm_id, yes=True)

src/vm-repair/azext_vm_repair/repair_utils.py

@@ -210,7 +210,7 @@ def _check_existing_rg(rg_name):
         logger.error(azCommandError)
         raise Exception('Unexpected error occured while fetching existing resource groups.')
 
-    logger.info('Resource group exists is \'%s\'', group_exists)
+    logger.info('Pre-existing repair resource group with the same name is \'%s\'', group_exists)
     return group_exists
 
 
@@ -491,22 +491,17 @@ def _fetch_compatible_windows_os_urn(source_vm):
 
 def _suse_image_selector(distro):
     fetch_urn_command = 'az vm image list --publisher SUSE --offer {offer} --sku gen1 --verbose --all --query "[].urn | reverse(sort(@))" -o json'.format(offer=distro)
-    logger.info('Fetching compatible SUSE OS images from gallery...')
     urns = loads(_call_az_command(fetch_urn_command))
 
     # Raise exception when not finding SUSE image
     if not urns:
         raise SuseNotAvailableError()
 
-    logger.debug('Fetched urns: \n%s', urns)
-    # Returning the first URN as it is the latest image with no special use like HPC or SAP
-    logger.debug('Return the first URN : %s', urns[0])
     return urns[0]
 
 
 def _suse_image_selector_gen2(distro):
     fetch_urn_command = 'az vm image list --publisher SUSE --offer {offer} --sku gen2 --verbose --all --query "[].urn | reverse(sort(@))" -o json'.format(offer=distro)
-    logger.info('Fetching compatible SUSE OS images from gallery...')
     urns = loads(_call_az_command(fetch_urn_command))
 
     # Raise exception when not finding SUSE image
@@ -711,6 +706,14 @@ def _unlock_encrypted_vm_run(repair_vm_name, repair_group_name, is_linux):
 
 
 def _create_repair_vm(copy_disk_id, create_repair_vm_command, repair_password, repair_username, fix_uuid=False):
+
+    # logging all parameters of the function individually
+    logger.info('Creating repair VM with command: {}'.format(create_repair_vm_command))
+    logger.info('copy_disk_id: {}'.format(copy_disk_id))
+    logger.info('repair_password: {}'.format(repair_password))
+    logger.info('repair_username: {}'.format(repair_username))
+    logger.info('fix_uuid: {}'.format(fix_uuid))
+
     if not fix_uuid:
         create_repair_vm_command += ' --attach-data-disks {id}'.format(id=copy_disk_id)
     logger.info('Validating VM template before continuing...')

src/vm-repair/azext_vm_repair/tests/latest/test_repair_commands.py

@@ -717,3 +717,28 @@ def test_vmrepair_ResetNicWindowsVM(self, resource_group):
         vm_instance_view = self.cmd('vm get-instance-view -g {rg} -n {vm} -o json').get_output_in_json()
         vm_power_state = vm_instance_view['instanceView']['statuses'][1]['code']
         assert vm_power_state == 'PowerState/running'
+
+
+@pytest.mark.repairandrestore
+class RepairAndRestoreLinuxVM(LiveScenarioTest):
+    
+    @ResourceGroupPreparer(location='westus2')
+    def test_vmrepair_RepairAndRestoreLinuxVM(self, resource_group):
+        self.kwargs.update({
+            'vm': 'vm1'
+        })
+
+        # Create test VM
+        self.cmd('vm create -g {rg} -n {vm} --admin-username azureadmin --image Win2016Datacenter --admin-password !Passw0rd2018')
+        vms = self.cmd('vm list -g {rg} -o json').get_output_in_json()
+        # Something wrong with vm create command if it fails here
+        assert len(vms) == 1
+
+        # Test Repair and restore
+        result = self.cmd('vm repair repair-and-restore -g {rg} -n {vm}')
+        assert result['status'] == STATUS_SUCCESS, result['error_message']
+
+        # Check swapped OS disk
+        vms = self.cmd('vm list -g {rg} -o json').get_output_in_json()
+        source_vm = vms[0]
+        assert source_vm['storageProfile']['osDisk']['name'] == result['copied_disk_name']

src/vm-repair/setup.py

@@ -8,7 +8,7 @@
 from codecs import open
 from setuptools import setup, find_packages
 
-VERSION = "0.5.3"
+VERSION = "0.5.4"
 
 CLASSIFIERS = [
     'Development Status :: 4 - Beta',