summaryrefslogtreecommitdiffstats
path: root/tests/functional/afr
diff options
context:
space:
mode:
authorVitalii Koriakov <vkoriako@redhat.com>2018-02-19 12:50:27 +0200
committerNigel Babu <nigelb@redhat.com>2018-04-20 07:12:44 +0000
commit139997b7e05e00a59e90767a2618a22a0284df81 (patch)
tree9a9257be7b1307fad1f8a2bf041ec5fe2281b066 /tests/functional/afr
parent62dbe9687e0cf5d076a72be967005aea9f3779e8 (diff)
Test data self-heal of files when "self-heal-algorithm" option value is "diff" (heal command)
Change-Id: Id310e0c17a872d8586ad8c7de79f1f68b93edb0a Signed-off-by: Vitalii Koriakov <vkoriako@redhat.com>
Diffstat (limited to 'tests/functional/afr')
-rwxr-xr-x[-rw-r--r--]tests/functional/afr/heal/test_self_heal.py204
1 files changed, 204 insertions, 0 deletions
diff --git a/tests/functional/afr/heal/test_self_heal.py b/tests/functional/afr/heal/test_self_heal.py
index b2e52e3..1ec4f01 100644..100755
--- a/tests/functional/afr/heal/test_self_heal.py
+++ b/tests/functional/afr/heal/test_self_heal.py
@@ -651,3 +651,207 @@ class TestSelfHeal(GlusterBaseClass):
'Checksums are not equal')
g.log.info('Checksums before bringing bricks online '
'and after bringing bricks online are equal')
+
+ def test_data_self_heal_algorithm_diff_heal_command(self):
+ """
+ Test Volume Option - 'cluster.data-self-heal-algorithm' : 'diff'
+
+ Description:
+ - set the volume option
+ "metadata-self-heal": "off"
+ "entry-self-heal": "off"
+ "data-self-heal": "off"
+ "data-self-heal-algorithm": "diff"
+ "self-heal-daemon": "off"
+ - create IO
+ - calculate areequal
+ - bring down all bricks processes from selected set
+ - modify the data
+ - get areequal before getting bricks online
+ - bring bricks online
+ - expand volume by adding bricks to the volume
+ - do rebalance
+ - set the volume option "self-heal-daemon": "on" and check for daemons
+ - start healing
+ - check if heal is completed
+ - check for split-brain
+ - calculate areequal and compare with arequal before bringing bricks
+ offline and after bringing bricks online
+ """
+ # pylint: disable=too-many-branches,too-many-statements
+ # Setting options
+ g.log.info('Setting options...')
+ options = {"metadata-self-heal": "off",
+ "entry-self-heal": "off",
+ "data-self-heal": "off",
+ "data-self-heal-algorithm": "diff"}
+ ret = set_volume_options(self.mnode, self.volname, options)
+ self.assertTrue(ret, 'Failed to set options')
+ g.log.info("Options "
+ "'metadata-self-heal', "
+ "'entry-self-heal', "
+ "'data-self-heal', "
+ "'self-heal-daemon' "
+ "are set to 'off',"
+ "'data-self-heal-algorithm' "
+ "is set to 'diff' successfully")
+
+ # Creating files on client side
+ for mount_object in self.mounts:
+ g.log.info("Generating data for %s:%s",
+ mount_object.client_system, mount_object.mountpoint)
+ # Creating files
+ command = ("python %s create_files -f 100 %s"
+ % (self.script_upload_path, mount_object.mountpoint))
+
+ proc = g.run_async(mount_object.client_system, command,
+ user=mount_object.user)
+ self.all_mounts_procs.append(proc)
+
+ # Validate IO
+ g.log.info("Wait for IO to complete and validate IO ...")
+ ret = validate_io_procs(self.all_mounts_procs, self.mounts)
+ self.io_validation_complete = True
+ self.assertTrue(ret, "IO failed on some of the clients")
+ g.log.info("IO is successful on all mounts")
+
+ # 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')
+ 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)
+
+ # Modify the data
+ self.all_mounts_procs = []
+ for mount_object in self.mounts:
+ g.log.info("Modifying data for %s:%s",
+ mount_object.client_system, mount_object.mountpoint)
+ command = ("python %s create_files -f 100 --fixed-file-size 1M %s"
+ % (self.script_upload_path, mount_object.mountpoint))
+
+ proc = g.run_async(mount_object.client_system, command,
+ user=mount_object.user)
+ self.all_mounts_procs.append(proc)
+
+ # Validate IO
+ g.log.info("Wait for IO to complete and validate IO ...")
+ ret = validate_io_procs(self.all_mounts_procs, self.mounts)
+ self.io_validation_complete = True
+ self.assertTrue(ret, "IO failed on some of the clients")
+ 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')
+
+ # 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)
+
+ # 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)
+
+ # Expand volume by adding bricks to the volume
+ g.log.info("Start adding bricks to volume...")
+ ret = expand_volume(self.mnode, self.volname, self.servers,
+ self.all_servers_info)
+ self.assertTrue(ret, ("Failed to expand the volume when IO in "
+ "progress on volume %s", self.volname))
+ g.log.info("Expanding volume is successful on volume %s", self.volname)
+ self.bricks_list = get_all_bricks(self.mnode, self.volname)
+
+ # Do rebalance
+ ret, _, _ = rebalance_start(self.mnode, self.volname)
+ self.assertEqual(ret, 0, 'Failed to start rebalance')
+ g.log.info('Rebalance is started')
+
+ ret = wait_for_rebalance_to_complete(self.mnode, self.volname)
+ self.assertTrue(ret, 'Rebalance is not completed')
+ g.log.info('Rebalance is completed successfully')
+
+ # 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')
+ g.log.info("Option 'self-heal-daemon' is set to 'on' successfully")
+
+ # 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')
+
+ # Checking areequals before bringing bricks offline
+ # and after bringing bricks online
+ self.assertItemsEqual(result_before_online, result_after_online,
+ 'Checksums are not equal')
+ g.log.info('Checksums are equal')