From 58813422c350068f9101bf9a890ab5753428ac7a Mon Sep 17 00:00:00 2001 From: Vitalii Koriakov Date: Fri, 16 Feb 2018 16:55:09 +0200 Subject: Test Entry-Self-Heal (heal command) Change-Id: Iaecdf6ad44677891340713a5c945a4bdc30ce527 Signed-off-by: Vitalii Koriakov --- tests/functional/afr/heal/test_self_heal.py | 251 ++++++++++++++++++++++++++++ 1 file changed, 251 insertions(+) (limited to 'tests/functional') diff --git a/tests/functional/afr/heal/test_self_heal.py b/tests/functional/afr/heal/test_self_heal.py index cf4ac6b17..7837d958c 100755 --- a/tests/functional/afr/heal/test_self_heal.py +++ b/tests/functional/afr/heal/test_self_heal.py @@ -414,3 +414,254 @@ class TestSelfHeal(GlusterBaseClass): self.assertTrue(ret, "IO failed on some of the clients") self.io_validation_complete = True g.log.info("IO is successful on all mounts") + + def test_entry_self_heal_heal_command(self): + """ + Test Entry-Self-Heal (heal command) + + Description: + - set the volume option + "metadata-self-heal": "off" + "entry-self-heal": "off" + "data-self-heal": "off" + - create IO + - get areequal before getting bricks offline + - set the volume option + "self-heal-daemon": "off" + - bring down all bricks processes from selected set + - get areequal after getting bricks offline and compare with + arequal after bringing bricks offline + - modify the data + - get areequal before getting bricks online + - bring bricks online + - set the volume option + "self-heal-daemon": "on" + - check daemons and start healing + - check if heal is completed + - check for split-brain + - get areequal after getting bricks online and compare with + arequal before bringing bricks online + """ + + # Setting options + g.log.info('Setting options...') + options = {"metadata-self-heal": "off", + "entry-self-heal": "off", + "data-self-heal": "off", + } + ret = set_volume_options(self.mnode, self.volname, options) + self.assertTrue(ret, 'Failed to set options %s' % options) + g.log.info("Options " + "'metadata-self-heal', " + "'entry-self-heal', " + "'data-self-heal', " + "are set to 'off'") + + # Start IO on mounts + g.log.info("Starting IO on all mounts...") + self.all_mounts_procs = [] + for mount_obj in self.mounts: + g.log.info("Starting IO on %s:%s" + % (mount_obj.client_system, + mount_obj.mountpoint)) + cmd = ("python %s create_deep_dirs_with_files " + "--dirname-start-num %d " + "--dir-length 2 " + "--dir-depth 2 " + "--max-num-of-dirs 2 " + "--num-of-files 20 %s" % (self.script_upload_path, + self.counter, + mount_obj.mountpoint)) + proc = g.run_async(mount_obj.client_system, cmd, + user=mount_obj.user) + self.all_mounts_procs.append(proc) + self.counter = self.counter + 10 + g.log.info("IO on %s:%s is started successfully" + % (mount_obj.client_system, + mount_obj.mountpoint)) + 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") + + # Command list to do different operations with data - + # create, rename, copy and delete + cmd_list = ["python %s create_files -f 20 %s", + "python %s mv -i '.trashcan' %s", + "python %s copy --dest-dir new_dir %s", + "python %s delete %s", + ] + + for cmd in cmd_list: + # Get areequal before getting bricks offline + g.log.info('Getting areequal before getting bricks offline...') + ret, result_before_offline = collect_mounts_arequal(self.mounts) + self.assertTrue(ret, 'Failed to get arequal') + g.log.info('Getting areequal before getting bricks offline ' + 'is successful') + + # Setting options + g.log.info('Setting options...') + options = {"self-heal-daemon": "off", + } + ret = set_volume_options(self.mnode, self.volname, options) + self.assertTrue(ret, 'Failed to set options %s' % options) + g.log.info("Option 'self-heal-daemon' " + "is set to 'off' successfully") + + # Select bricks to bring offline + bricks_to_bring_offline_dict = (select_bricks_to_bring_offline( + self.mnode, self.volname)) + bricks_to_bring_offline = filter(None, ( + bricks_to_bring_offline_dict['hot_tier_bricks'] + + bricks_to_bring_offline_dict['cold_tier_bricks'] + + bricks_to_bring_offline_dict['volume_bricks'])) + + # Bring brick offline + 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) + + # Get areequal after getting bricks offline + g.log.info('Getting areequal after getting bricks offline...') + ret, result_after_offline = collect_mounts_arequal(self.mounts) + self.assertTrue(ret, 'Failed to get arequal') + g.log.info('Getting areequal after getting bricks offline ' + 'is successful') + + # Checking areequals before bringing bricks offline + # and after bringing bricks offline + self.assertEqual(result_before_offline, result_after_offline, + 'Checksums are not equal') + g.log.info('Checksums before bringing bricks offline ' + 'and after bringing bricks offline are equal') + + # Modify the data + g.log.info("Start modifying IO on all mounts...") + self.all_mounts_procs = [] + for mount_obj in self.mounts: + g.log.info("Modifying IO on %s:%s", mount_obj.client_system, + mount_obj.mountpoint) + cmd = cmd % (self.script_upload_path, mount_obj.mountpoint) + proc = g.run_async(mount_obj.client_system, cmd, + user=mount_obj.user) + self.all_mounts_procs.append(proc) + g.log.info("IO on %s:%s is modified successfully" + % (mount_obj.client_system, + mount_obj.mountpoint)) + 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 areequal before getting bricks online + g.log.info('Getting areequal before getting bricks online...') + ret, result_before_online = collect_mounts_arequal(self.mounts) + self.assertTrue(ret, 'Failed to get arequal') + g.log.info('Getting areequal before getting bricks online ' + 'is successful') + + # 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") + + # Bring 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) + + # Setting options + g.log.info('Setting options...') + options = {"self-heal-daemon": "on", + } + ret = set_volume_options(self.mnode, self.volname, options) + self.assertTrue(ret, 'Failed to set options %s' % options) + g.log.info("Option 'self-heal-daemon' is set to 'on' successfully") + + # Wait for volume processes to be online + g.log.info("Wait for volume processes to be online") + ret = wait_for_volume_process_to_be_online(self.mnode, + self.volname) + self.assertTrue(ret, ("Failed to wait for volume %s processes to " + "be online", self.volname)) + g.log.info("Successful in waiting for volume %s processes to be " + "online", self.volname) + + # Verify volume's all process are online + g.log.info("Verifying volume's all process are online") + ret = verify_all_process_of_volume_are_online(self.mnode, + self.volname) + self.assertTrue(ret, ("Volume %s : All process are not online" + % self.volname)) + g.log.info("Volume %s : All process are online" % self.volname) + + # Wait for self-heal-daemons to be online + g.log.info("Waiting for self-heal-daemons to be online") + ret = is_shd_daemonized(self.all_servers) + self.assertTrue(ret, "Either No self heal daemon process found") + g.log.info("All self-heal-daemons are online") + + # 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') + + # Get areequal after getting bricks online + g.log.info('Getting areequal after getting bricks online...') + ret, result_after_online = collect_mounts_arequal(self.mounts) + self.assertTrue(ret, 'Failed to get arequal') + g.log.info('Getting areequal after getting bricks online ' + 'is successful') + + # 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") + + # Checking areequals before bringing bricks online + # and after bringing bricks online + self.assertEqual(result_before_online, result_after_online, + 'Checksums are not equal') + g.log.info('Checksums before bringing bricks online ' + 'and after bringing bricks online are equal') -- cgit