Example #1
0
    def setUpClass(cls):
        """Initialize all the variables necessary for test cases."""
        super(BaseClass, cls).setUpClass()

        # Initializes OCP config variables
        cls.ocp_servers_info = g.config['ocp_servers']
        cls.ocp_master_node = g.config['ocp_servers']['master'].keys()
        cls.ocp_master_node_info = g.config['ocp_servers']['master']
        cls.ocp_client = g.config['ocp_servers']['client'].keys()
        cls.ocp_client_info = g.config['ocp_servers']['client']
        cls.ocp_nodes = g.config['ocp_servers']['nodes'].keys()
        cls.ocp_nodes_info = g.config['ocp_servers']['nodes']

        # Initializes storage project config variables
        openshift_config = g.config.get("cns", g.config.get("openshift"))
        cls.storage_project_name = openshift_config.get(
            'storage_project_name',
            openshift_config.get('setup', {}).get('cns_project_name'))

        # Initializes heketi config variables
        heketi_config = openshift_config['heketi_config']
        cls.heketi_dc_name = heketi_config['heketi_dc_name']
        cls.heketi_service_name = heketi_config['heketi_service_name']
        cls.heketi_client_node = heketi_config['heketi_client_node']
        cls.heketi_server_url = heketi_config['heketi_server_url']
        cls.heketi_cli_user = heketi_config['heketi_cli_user']
        cls.heketi_cli_key = heketi_config['heketi_cli_key']

        cls.gluster_servers = g.config['gluster_servers'].keys()
        cls.gluster_servers_info = g.config['gluster_servers']

        cls.storage_classes = openshift_config['dynamic_provisioning'][
            'storage_classes']
        cls.sc = cls.storage_classes.get(
            'storage_class1', cls.storage_classes.get('file_storage_class'))
        cmd = "echo -n %s | base64" % cls.heketi_cli_key
        ret, out, err = g.run(cls.ocp_master_node[0], cmd, "root")
        if ret != 0:
            raise ExecutionError("failed to execute cmd %s on %s out: %s "
                                 "err: %s" %
                                 (cmd, cls.ocp_master_node[0], out, err))
        cls.secret_data_key = out.strip()

        # Checks if heketi server is alive
        if not hello_heketi(cls.heketi_client_node, cls.heketi_server_url):
            raise ConfigError("Heketi server %s is not alive" %
                              cls.heketi_server_url)

        # Switch to the storage project
        if not switch_oc_project(cls.ocp_master_node[0],
                                 cls.storage_project_name):
            raise ExecutionError("Failed to switch oc project on node %s" %
                                 cls.ocp_master_node[0])

        if 'glustotest_run_id' not in g.config:
            g.config['glustotest_run_id'] = (
                datetime.datetime.now().strftime('%H_%M_%d_%m_%Y'))
        cls.glustotest_run_id = g.config['glustotest_run_id']
        msg = "Setupclass: %s : %s" % (cls.__name__, cls.glustotest_run_id)
        g.log.info(msg)
    def test_delete_heketidb_volume(self):
        """
        Method to test heketidb volume deletion via heketi-cli
        """
        volume_id_list = []
        heketidbexists = False
        msg = "Error: Cannot delete volume containing the Heketi database"

        for i in range(0, 2):
            volume_info = heketi_ops.heketi_volume_create(
                self.heketi_client_node, self.heketi_server_url, 10, json=True)
            self.assertNotEqual(volume_info, False, "Volume creation failed")
            volume_id_list.append(volume_info["id"])

        self.addCleanup(self.delete_volumes, volume_id_list)

        volume_list_info = heketi_ops.heketi_volume_list(
            self.heketi_client_node, self.heketi_server_url, json=True)

        self.assertNotEqual(volume_list_info, False,
                            "Heketi volume list command failed")

        if volume_list_info["volumes"] == []:
            raise ExecutionError("Heketi volume list empty")

        for volume_id in volume_list_info["volumes"]:
            volume_info = heketi_ops.heketi_volume_info(
                self.heketi_client_node,
                self.heketi_server_url,
                volume_id,
                json=True)

            if volume_info["name"] == "heketidbstorage":
                heketidbexists = True
                delete_ret, delete_output, delete_error = (
                    heketi_ops.heketi_volume_delete(self.heketi_client_node,
                                                    self.heketi_server_url,
                                                    volume_id,
                                                    raw_cli_output=True))

                self.assertNotEqual(delete_ret, 0, "Return code not 0")
                self.assertEqual(
                    delete_error.strip(), msg,
                    "Invalid reason for heketidb deletion failure")

        if not heketidbexists:
            raise ExecutionError(
                "Warning: heketidbstorage doesn't exist in list of volumes")
Example #3
0
    def test_dynamic_provisioning_glusterfile_glusterpod_failure(self):
        """Validate dynamic provisioning for gluster file when gluster pod down
        """
        mount_path = "/mnt"
        datafile_path = '%s/fake_file_for_%s' % (mount_path, self.id())

        # Create secret and storage class
        self.create_storage_class()

        # Create PVC
        pvc_name = self.create_and_wait_for_pvc()

        # Create app POD with attached volume
        pod_name = oc_create_tiny_pod_with_volume(
            self.node, pvc_name, "test-pvc-mount-on-app-pod",
            mount_path=mount_path)
        self.addCleanup(
            wait_for_resource_absence, self.node, 'pod', pod_name)
        self.addCleanup(oc_delete, self.node, 'pod', pod_name)

        # Wait for app POD be up and running
        wait_for_pod_be_ready(
            self.node, pod_name, timeout=60, wait_step=2)

        # Run IO in background
        io_cmd = "oc rsh %s dd if=/dev/urandom of=%s bs=1000K count=900" % (
            pod_name, datafile_path)
        async_io = g.run_async(self.node, io_cmd, "root")

        # Pick up one of the hosts which stores PV brick (4+ nodes case)
        gluster_pod_data = get_gluster_pod_names_by_pvc_name(
            self.node, pvc_name)[0]

        # Delete glusterfs POD from chosen host and wait for spawn of new one
        oc_delete(self.node, 'pod', gluster_pod_data["pod_name"])
        cmd = ("oc get pods -o wide | grep glusterfs | grep %s | "
               "grep -v Terminating | awk '{print $1}'") % (
                   gluster_pod_data["host_name"])
        for w in Waiter(600, 30):
            out = self.cmd_run(cmd)
            new_gluster_pod_name = out.strip().split("\n")[0].strip()
            if not new_gluster_pod_name:
                continue
            else:
                break
        if w.expired:
            error_msg = "exceeded timeout, new gluster pod not created"
            g.log.error(error_msg)
            raise ExecutionError(error_msg)
        new_gluster_pod_name = out.strip().split("\n")[0].strip()
        g.log.info("new gluster pod name is %s" % new_gluster_pod_name)
        wait_for_pod_be_ready(self.node, new_gluster_pod_name)

        # Check that async IO was not interrupted
        ret, out, err = async_io.async_communicate()
        self.assertEqual(ret, 0, "IO %s failed on %s" % (io_cmd, self.node))
Example #4
0
    def _node_reboot(self):
        storage_hostname = (
            g.config["gluster_servers"][self.gluster_servers[0]]["storage"])

        cmd = "sleep 3; /sbin/shutdown -r now 'Reboot triggered by Glusto'"
        ret, out, err = g.run(storage_hostname, cmd)

        self.addCleanup(self._wait_for_gluster_pod_to_be_ready)

        if ret != 255:
            err_msg = "failed to reboot host %s error: %s" % (storage_hostname,
                                                              err)
            g.log.error(err_msg)
            raise AssertionError(err_msg)

        try:
            g.ssh_close_connection(storage_hostname)
        except Exception as e:
            g.log.error("failed to close connection with host %s"
                        " with error: %s" % (storage_hostname, e))
            raise

        # added sleep as node will restart after 3 sec
        time.sleep(3)

        for w in Waiter(timeout=600, interval=10):
            try:
                if g.rpyc_get_connection(storage_hostname, user="******"):
                    g.rpyc_close_connection(storage_hostname, user="******")
                    break
            except Exception as err:
                g.log.info("exception while getting connection: '%s'" % err)

        if w.expired:
            error_msg = ("exceeded timeout 600 sec, node '%s' is "
                         "not reachable" % storage_hostname)
            g.log.error(error_msg)
            raise ExecutionError(error_msg)

        # wait for the gluster pod to be in 'Running' state
        self._wait_for_gluster_pod_to_be_ready()

        # glusterd and gluster-blockd service should be up and running
        service_names = ("glusterd", "gluster-blockd", "tcmu-runner")
        for gluster_pod in self.gluster_pod_list:
            for service in service_names:
                g.log.info("gluster_pod - '%s' : gluster_service '%s'" %
                           (gluster_pod, service))
                check_service_status(self.oc_node, gluster_pod, service,
                                     "running")
    def setUpClass(cls):
        """
        setUpClass of HeketiBaseClass
        """

        super(HeketiBaseClass, cls).setUpClass()

        # Initializes config variables
        openshift_config = g.config.get("cns", g.config.get("openshift"))
        cls.storage_project_name = openshift_config.get(
            'storage_project_name',
            openshift_config.get('setup', {}).get('cns_project_name'))

        cls.ocp_master_nodes = g.config['ocp_servers']['master'].keys()
        cls.ocp_master_node = cls.ocp_master_nodes[0]

        heketi_config = openshift_config['heketi_config']
        cls.heketi_dc_name = heketi_config['heketi_dc_name']
        cls.heketi_service_name = heketi_config['heketi_service_name']
        cls.heketi_client_node = heketi_config['heketi_client_node']
        cls.heketi_server_url = heketi_config['heketi_server_url']
        cls.heketi_cli_user = heketi_config['heketi_cli_user']
        cls.heketi_cli_key = heketi_config['heketi_cli_key']
        cls.gluster_servers = g.config['gluster_servers'].keys()
        cls.gluster_servers_info = g.config['gluster_servers']

        # Checks if heketi server is alive
        if not hello_heketi(cls.heketi_client_node, cls.heketi_server_url):
            raise ConfigError("Heketi server %s is not alive" %
                              cls.heketi_server_url)

        # Switch to the storage project
        if not openshift_ops.switch_oc_project(cls.ocp_master_node,
                                               cls.storage_project_name):
            raise ExecutionError("Failed to switch oc project on node %s" %
                                 cls.ocp_master_node)

        # Have a unique string to recognize the test run for logging
        if 'glustotest_run_id' not in g.config:
            g.config['glustotest_run_id'] = (
                datetime.datetime.now().strftime('%H_%M_%d_%m_%Y'))
        cls.glustotest_run_id = g.config['glustotest_run_id']
        msg = "Setupclass: %s : %s" % (cls.__name__, cls.glustotest_run_id)
        g.log.info(msg)
Example #6
0
    def _wait_for_gluster_pod_to_be_ready(self):
        for gluster_pod in self.gluster_pod_list:
            for w in Waiter(timeout=600, interval=10):
                try:
                    success = wait_for_pod_be_ready(self.oc_node,
                                                    gluster_pod,
                                                    timeout=1,
                                                    wait_step=1)
                    if success:
                        break
                except ExecutionError as e:
                    g.log.info("exception %s while validating gluster "
                               "pod %s" % (e, gluster_pod))

            if w.expired:
                error_msg = ("exceeded timeout 600 sec, pod '%s' is "
                             "not in 'running' state" % gluster_pod)
                g.log.error(error_msg)
                raise ExecutionError(error_msg)
    def delete_volumes(self, volume_ids):
        """
        Delete volumes by their IDs and raise error with list of failures
        Input: (volume_ids) It can be a single volume ID
        or a list of volume IDs
        """
        errored_ids = []

        if not isinstance(volume_ids, (list, set, tuple)):
            volume_ids = [volume_ids]

        for volume_id in volume_ids:
            out = heketi_volume_delete(self.heketi_client_node,
                                       self.heketi_server_url, volume_id)
            output_str = 'Volume %s deleted' % volume_id
            if output_str not in out:
                errored_ids.append(volume_id)

        if errored_ids:
            raise ExecutionError("Failed to delete following heketi volumes: "
                                 "%s" % ',\n'.join(errored_ids))
def enable_pvc_resize(master_node):
    '''
     This function edits the /etc/origin/master/master-config.yaml
     file - to enable pv_resize feature
     and restarts atomic-openshift service on master node
     Args:
         master_node (str): hostname of masternode  on which
                           want to edit the
                           master-config.yaml file
     Returns:
         bool: True if successful,
               otherwise raise Exception
    '''
    version = get_openshift_version()
    if version < "3.9":
        msg = ("pv resize is not available in openshift "
               "version %s " % version)
        g.log.error(msg)
        raise NotSupportedException(msg)

    try:
        conn = g.rpyc_get_connection(master_node, user="******")
        if conn is None:
            err_msg = ("Failed to get rpyc connection of node %s" %
                       master_node)
            g.log.error(err_msg)
            raise ExecutionError(err_msg)

        with conn.builtin.open(MASTER_CONFIG_FILEPATH, 'r') as f:
            data = yaml.load(f)
            dict_add = data['admissionConfig']['pluginConfig']
            if "PersistentVolumeClaimResize" in dict_add:
                g.log.info("master-config.yaml file is already edited")
                return True
            dict_add['PersistentVolumeClaimResize'] = {
                'configuration': {
                    'apiVersion': 'v1',
                    'disable': 'false',
                    'kind': 'DefaultAdmissionConfig'
                }
            }
            data['admissionConfig']['pluginConfig'] = dict_add
            kube_config = data['kubernetesMasterConfig']
            for key in ('apiServerArguments', 'controllerArguments'):
                kube_config[key] = (kube_config.get(key) if isinstance(
                    kube_config.get(key), dict) else {})
                value = ['ExpandPersistentVolumes=true']
                kube_config[key]['feature-gates'] = value
        with conn.builtin.open(MASTER_CONFIG_FILEPATH, 'w+') as f:
            yaml.dump(data, f, default_flow_style=False)
    except Exception as err:
        raise ExecutionError("failed to edit master-config.yaml file "
                             "%s on %s" % (err, master_node))
    finally:
        g.rpyc_close_connection(master_node, user="******")

    g.log.info("successfully edited master-config.yaml file "
               "%s" % master_node)
    if version == "3.9":
        cmd = ("systemctl restart atomic-openshift-master-api "
               "atomic-openshift-master-controllers")
    else:
        cmd = ("/usr/local/bin/master-restart api && "
               "/usr/local/bin/master-restart controllers")
    ret, out, err = g.run(master_node, cmd, "root")
    if ret != 0:
        err_msg = "Failed to execute cmd %s on %s\nout: %s\nerr: %s" % (
            cmd, master_node, out, err)
        g.log.error(err_msg)
        raise ExecutionError(err_msg)

    return True