diff options
Diffstat (limited to 'tests')
-rw-r--r-- | tests/functional/glusterd/test_volume_set_when_glusterd_stopped_on_one_node.py | 193 |
1 files changed, 193 insertions, 0 deletions
diff --git a/tests/functional/glusterd/test_volume_set_when_glusterd_stopped_on_one_node.py b/tests/functional/glusterd/test_volume_set_when_glusterd_stopped_on_one_node.py new file mode 100644 index 000000000..d99fa185f --- /dev/null +++ b/tests/functional/glusterd/test_volume_set_when_glusterd_stopped_on_one_node.py @@ -0,0 +1,193 @@ +# Copyright (C) 2020 Red Hat, Inc. <http://www.redhat.com> +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License along +# with this program; if not, write to the Free Software Foundation, Inc., +# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + +""" Description: + Volume set operation when glusterd is stopped on one node +""" + +from random import choice +from time import sleep +from glusto.core import Glusto as g +from glustolibs.gluster.gluster_base_class import GlusterBaseClass, runs_on +from glustolibs.gluster.exceptions import ExecutionError +from glustolibs.gluster.volume_ops import ( + set_volume_options, get_volume_info) +from glustolibs.gluster.brick_libs import get_online_bricks_list +from glustolibs.gluster.gluster_init import ( + start_glusterd, stop_glusterd, wait_for_glusterd_to_start) +from glustolibs.misc.misc_libs import upload_scripts +from glustolibs.io.utils import validate_io_procs + + +@runs_on([['distributed', 'replicated', 'distributed-replicated', + 'dispersed', 'distributed-dispersed'], ['glusterfs']]) +class TestVolumeSetWhenGlusterdStoppedOnOneNode(GlusterBaseClass): + + @classmethod + def setUpClass(cls): + cls.get_super_method(cls, 'setUpClass')() + + # Uploading file_dir script in all client direcotries + cls.script_upload_path = ("/usr/share/glustolibs/io/scripts/" + "file_dir_ops.py") + ret = upload_scripts(cls.clients, cls.script_upload_path) + if not ret: + raise ExecutionError("Failed to upload IO scripts to clients %s" + % cls.clients) + g.log.info("Successfully uploaded IO scripts to clients %s", + cls.clients) + + def setUp(self): + self.get_super_method(self, 'setUp')() + # Creating Volume and mounting volume. + ret = self.setup_volume_and_mount_volume(self.mounts) + if not ret: + raise ExecutionError("Volume creation or mount failed: %s" + % self.volname) + g.log.info("Volme created and mounted successfully : %s", + self.volname) + + def tearDown(self): + # Check if a node is still down + if self.glusterd_is_stopped: + ret = start_glusterd(self.random_server) + self.assertTrue(ret, "Failed to start glusterd on %s" + % self.random_server) + g.log.info("Successfully started glusterd on node: %s", + self.random_server) + + # Waiting for glusterd to start completely + ret = wait_for_glusterd_to_start(self.random_server) + self.assertTrue(ret, "glusterd is not running on %s" + % self.random_server) + g.log.info("glusterd is started and running on %s", + self.random_server) + + # Unmounting and cleaning volume. + ret = self.unmount_volume_and_cleanup_volume(self.mounts) + if not ret: + raise ExecutionError("Unable to delete volume % s" % self.volname) + g.log.info("Volume deleted successfully : %s", self.volname) + + self.get_super_method(self, 'tearDown')() + + def test_volume_set_when_glusterd_stopped_on_one_node(self): + """ + Test Case: + 1) Setup and mount a volume on client. + 2) Stop glusterd on a random server. + 3) Start IO on mount points + 4) Set an option on the volume + 5) Start glusterd on the stopped node. + 6) Verify all the bricks are online after starting glusterd. + 7) Check if the volume info is synced across the cluster. + """ + # Fetching the bricks list and storing it for later use + list1 = get_online_bricks_list(self.mnode, self.volname) + self.assertIsNotNone(list1, "Failed to get the list of online bricks " + "for volume: %s" % self.volname) + + # Fetching a random server from list. + self.random_server = choice(self.servers[1:]) + + # Stopping glusterd on one node. + ret = stop_glusterd(self.random_server) + self.assertTrue(ret, "Failed to stop glusterd on one node.") + g.log.info("Successfully stopped glusterd on one node.") + + self.glusterd_is_stopped = True + + # Start IO on mount points. + self.all_mounts_procs = [] + counter = 1 + for mount_obj in self.mounts: + g.log.info("Starting IO on %s:%s", mount_obj.client_system, + mount_obj.mountpoint) + cmd = ("/usr/bin/env python %s create_deep_dirs_with_files " + "--dir-depth 4 " + "--dir-length 6 " + "--dirname-start-num %d " + "--max-num-of-dirs 3 " + "--num-of-files 5 %s" % ( + self.script_upload_path, + counter, mount_obj.mountpoint)) + proc = g.run_async(mount_obj.client_system, cmd, + user=mount_obj.user) + self.all_mounts_procs.append(proc) + counter += 1 + + # Validate IO + self.assertTrue( + validate_io_procs(self.all_mounts_procs, self.mounts), + "IO failed on some of the clients" + ) + g.log.info("IO validation complete.") + + # set a option on volume, stat-prefetch on + self.options = {"stat-prefetch": "on"} + ret = set_volume_options(self.mnode, self.volname, self.options) + self.assertTrue(ret, ("Failed to set option stat-prefetch to on" + "for the volume %s" % self.volname)) + g.log.info("Succeeded in setting stat-prefetch option to on" + "for the volume %s", self.volname) + + # start glusterd on the node where glusterd is stopped + ret = start_glusterd(self.random_server) + self.assertTrue(ret, "Failed to start glusterd on %s" + % self.random_server) + g.log.info("Successfully started glusterd on node: %s", + self.random_server) + + # Waiting for glusterd to start completely + ret = wait_for_glusterd_to_start(self.random_server) + self.assertTrue(ret, "glusterd is not running on %s" + % self.random_server) + g.log.info("glusterd is started and running on %s", self.random_server) + + self.glusterd_is_stopped = False + + # Confirm if all the bricks are online or not + count = 0 + while count < 10: + list2 = get_online_bricks_list(self.mnode, self.volname) + if list1 == list2: + break + sleep(2) + count += 1 + + self.assertListEqual(list1, list2, "Unexpected: All the bricks in the" + "volume are not online") + g.log.info("All the bricks in the volume are back online") + + # volume info should be synced across the cluster + out1 = get_volume_info(self.mnode, self.volname) + self.assertIsNotNone(out1, "Failed to get the volume info from %s" + % self.mnode) + g.log.info("Getting volume info from %s is success", self.mnode) + + count = 0 + while count < 60: + out2 = get_volume_info(self.random_server, self.volname) + self.assertIsNotNone(out2, "Failed to get the volume info from %s" + % self.random_server) + if out1 == out2: + break + sleep(2) + count += 1 + + self.assertDictEqual(out1, out2, "Volume info is not synced in the" + "restarted node") + g.log.info("Volume info is successfully synced across the cluster") |