From b585613e73ef8a2906e21d03e2050a4308c5b772 Mon Sep 17 00:00:00 2001 From: Vitalii Koriakov Date: Fri, 15 Jun 2018 14:41:21 +0300 Subject: Resolving meta-data split-brain from client using extended attributes Change-Id: I2ba674b8ea97964040f2e7d47a169c1e41808116 Signed-off-by: Vitalii Koriakov --- ...ng_meta_data_split_brain_extended_attributes.py | 257 +++++++++++++++++++++ 1 file changed, 257 insertions(+) create mode 100644 tests/functional/arbiter/test_resolving_meta_data_split_brain_extended_attributes.py (limited to 'tests') diff --git a/tests/functional/arbiter/test_resolving_meta_data_split_brain_extended_attributes.py b/tests/functional/arbiter/test_resolving_meta_data_split_brain_extended_attributes.py new file mode 100644 index 000000000..e6f68c0cf --- /dev/null +++ b/tests/functional/arbiter/test_resolving_meta_data_split_brain_extended_attributes.py @@ -0,0 +1,257 @@ +# Copyright (C) 2015-2018 Red Hat, Inc. +# +# 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. + +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.heal_ops import trigger_heal +from glustolibs.gluster.volume_libs import get_subvols +from glustolibs.gluster.brick_libs import (bring_bricks_offline, + bring_bricks_online, + are_bricks_offline) +from glustolibs.gluster.heal_libs import (monitor_heal_completion, + is_heal_complete, + is_volume_in_split_brain) +from glustolibs.io.utils import (validate_io_procs, + list_all_files_and_dirs_mounts, + wait_for_io_to_complete) + + +@runs_on([['distributed-replicated'], + ['glusterfs', 'nfs', 'cifs']]) +class TestArbiterSelfHeal(GlusterBaseClass): + """ + Description: + Arbiter Test cases related to + healing in default configuration of the volume + """ + + def setUp(self): + # Calling GlusterBaseClass setUp + GlusterBaseClass.setUp.im_func(self) + + # Setup Volumes + if self.volume_type == "distributed-replicated": + # Redefine distributed-replicated volume + self.volume['voltype'] = { + 'type': 'distributed-replicated', + 'replica_count': 3, + 'dist_count': 2, + 'arbiter_count': 1, + 'transport': 'tcp'} + + self.all_mounts_procs = [] + self.io_validation_complete = False + + # Setup Volume and Mount Volume + g.log.info("Starting to Setup Volume and Mount Volume") + ret = self.setup_volume_and_mount_volume(mounts=self.mounts) + if not ret: + raise ExecutionError("Failed to Setup_Volume and Mount_Volume") + g.log.info("Successful in Setup Volume and Mount Volume") + + def tearDown(self): + """ + If test method failed before validating IO, tearDown waits for the + IO's to complete and checks for the IO exit status + + Cleanup and umount volume + """ + if not self.io_validation_complete: + g.log.info("Wait for IO to complete as IO validation did not " + "succeed in test method") + ret = wait_for_io_to_complete(self.all_mounts_procs, self.mounts) + if not ret: + raise ExecutionError("IO failed on some of the clients") + g.log.info("IO is successful on all mounts") + + # List all files and dirs created + g.log.info("List all files and directories:") + ret = list_all_files_and_dirs_mounts(self.mounts) + if not ret: + raise ExecutionError("Failed to list all files and dirs") + g.log.info("Listing all files and directories is successful") + + # Cleanup and umount volume + g.log.info("Starting to Unmount Volume and Cleanup Volume") + ret = self.unmount_volume_and_cleanup_volume(mounts=self.mounts) + if not ret: + raise ExecutionError("Failed to umount the vol & cleanup Volume") + g.log.info("Successful in umounting the volume and Cleanup") + + # Calling GlusterBaseClass teardown + GlusterBaseClass.tearDown.im_func(self) + + def test_resolving_meta_data(self): + """ + - Create a file test_file.txt + - Find out which brick the file resides on and kill arbiter brick + in the replica pair + - Modify the permissions of the file + - Bring back the killed brick + - Kill the other brick in the replica pair + - Modify the permissions of the file + - Bring back the killed brick + - Trigger heal + - Check if heal is completed + - Check for split-brain + """ + # pylint: disable=too-many-locals,too-many-statements + # Creating files on client side + file_to_create = 'test_file.txt' + for mount_obj in self.mounts: + g.log.info("Generating data for %s:%s", + mount_obj.client_system, mount_obj.mountpoint) + # Create file + g.log.info('Creating file...') + command = ("cd %s ; " + "touch %s" + % (mount_obj.mountpoint, file_to_create)) + + proc = g.run_async(mount_obj.client_system, command, + user=mount_obj.user) + self.all_mounts_procs.append(proc) + self.io_validation_complete = False + + # Validate IO + g.log.info("Wait for IO to complete and validate IO ...") + ret = validate_io_procs(self.all_mounts_procs, self.mounts) + self.assertTrue(ret, "IO failed on some of the clients") + self.io_validation_complete = True + g.log.info("IO is successful on all mounts") + + # get bricks with file + g.log.info('Getting bricks with file...') + subvols_dict = get_subvols(self.mnode, self.volname) + brick_list_with_file = [] + for subvol in subvols_dict['volume_subvols']: + for brick in subvol: + node, brick_path = brick.split(':') + ret, brick_file_list, _ = g.run(node, 'ls %s' % brick_path) + if 'test_file.txt' in brick_file_list: + brick_list_with_file.append(brick) + g.log.info('Bricks with file: %s', brick_list_with_file) + + # Bring arbiter brick offline + bricks_to_bring_offline = [brick_list_with_file[-1]] + g.log.info('Bringing bricks %s offline...', bricks_to_bring_offline) + ret = bring_bricks_offline(self.volname, bricks_to_bring_offline) + self.assertTrue(ret, 'Failed to bring bricks %s offline' % + bricks_to_bring_offline) + + ret = are_bricks_offline(self.mnode, self.volname, + bricks_to_bring_offline) + self.assertTrue(ret, 'Bricks %s are not offline' + % bricks_to_bring_offline) + g.log.info('Bringing bricks %s offline is successful', + bricks_to_bring_offline) + + # Modify the data + self.all_mounts_procs = [] + for mount_obj in self.mounts: + g.log.info("Modifying data for %s:%s", + mount_obj.client_system, mount_obj.mountpoint) + # Modify the permissions + g.log.info('Modifying the permissions of the file...') + command = ("cd %s ; " + "chmod 600 %s" + % (mount_obj.mountpoint, file_to_create)) + + proc = g.run_async(mount_obj.client_system, command, + user=mount_obj.user) + self.all_mounts_procs.append(proc) + self.io_validation_complete = False + + # Validate IO + g.log.info("Wait for IO to complete and validate IO ...") + ret = validate_io_procs(self.all_mounts_procs, self.mounts) + self.assertTrue(ret, "IO failed on some of the clients") + self.io_validation_complete = True + g.log.info("IO is successful on all mounts") + + # Bring arbiter brick online + g.log.info('Bringing bricks %s online...', bricks_to_bring_offline) + ret = bring_bricks_online(self.mnode, self.volname, + bricks_to_bring_offline) + self.assertTrue(ret, 'Failed to bring bricks %s online' % + bricks_to_bring_offline) + g.log.info('Bringing bricks %s online is successful', + bricks_to_bring_offline) + + # Bring 1-st data brick offline + bricks_to_bring_offline = [brick_list_with_file[0]] + g.log.info('Bringing bricks %s offline...', bricks_to_bring_offline) + ret = bring_bricks_offline(self.volname, bricks_to_bring_offline) + self.assertTrue(ret, 'Failed to bring bricks %s offline' % + bricks_to_bring_offline) + + ret = are_bricks_offline(self.mnode, self.volname, + bricks_to_bring_offline) + self.assertTrue(ret, 'Bricks %s are not offline' + % bricks_to_bring_offline) + g.log.info('Bringing bricks %s offline is successful', + bricks_to_bring_offline) + + # Modify the data + self.all_mounts_procs = [] + for mount_obj in self.mounts: + g.log.info("Modifying data for %s:%s", + mount_obj.client_system, mount_obj.mountpoint) + # Modify the permissions + g.log.info('Modifying the permissions of the file...') + command = ("cd %s ; " + "chmod 644 %s" + % (mount_obj.mountpoint, file_to_create)) + + proc = g.run_async(mount_obj.client_system, command, + user=mount_obj.user) + self.all_mounts_procs.append(proc) + self.io_validation_complete = False + + # Validate IO + g.log.info("Wait for IO to complete and validate IO ...") + ret = validate_io_procs(self.all_mounts_procs, self.mounts) + self.assertTrue(ret, "IO failed on some of the clients") + self.io_validation_complete = True + g.log.info("IO is successful on all mounts") + + # Bring 1-st data brick online + g.log.info('Bringing bricks %s online...', bricks_to_bring_offline) + ret = bring_bricks_online(self.mnode, self.volname, + bricks_to_bring_offline) + self.assertTrue(ret, 'Failed to bring bricks %s online' % + bricks_to_bring_offline) + g.log.info('Bringing bricks %s online is successful', + bricks_to_bring_offline) + + # Start healing + ret = trigger_heal(self.mnode, self.volname) + self.assertTrue(ret, 'Heal is not started') + g.log.info('Healing is started') + + # Monitor heal completion + ret = monitor_heal_completion(self.mnode, self.volname) + self.assertTrue(ret, 'Heal has not yet completed') + + # Check if heal is completed + ret = is_heal_complete(self.mnode, self.volname) + self.assertTrue(ret, 'Heal is not complete') + g.log.info('Heal is completed successfully') + + # Check for split-brain + ret = is_volume_in_split_brain(self.mnode, self.volname) + self.assertFalse(ret, 'Volume is in split-brain state') + g.log.info('Volume is not in split-brain state') -- cgit