-
Notifications
You must be signed in to change notification settings - Fork 93
/
longevity_alternator_ttl_test.py
174 lines (143 loc) · 7.91 KB
/
longevity_alternator_ttl_test.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
import time
from longevity_test import LongevityTest
from sdcm.cluster import BaseNode
from sdcm.utils.alternator.consts import NO_LWT_TABLE_NAME
class AlternatorTtlLongevityTest(LongevityTest):
keyspace = f"alternator_{NO_LWT_TABLE_NAME}"
full_table_name = f'{keyspace}.{NO_LWT_TABLE_NAME}'
def _count_sstables_and_partitions(self, node: BaseNode) -> int:
cfstats = node.get_cfstats(self.full_table_name)
estimated_num_of_partitions = cfstats['Number of partitions (estimate)']
num_of_sstables = cfstats['SSTable count']
self.log.info('Table stats results are: %s sstables, %s estimated partitions',
num_of_sstables, estimated_num_of_partitions)
self.log.info('Run a table scan to count number of existing items')
with self.db_cluster.cql_connection_patient(node=node, connect_timeout=600) as session:
result = session.execute(f"SELECT count(*) FROM {self.full_table_name} using timeout 10m")
partitions_num_result = result.current_rows[0].count
self.log.info('Number of partitions found: %s', partitions_num_result)
return partitions_num_result
def count_sstables_and_partitions_after_major_compaction(self, run_repair: bool = True):
if run_repair:
self.log.info('Run a repair on nodes..')
for node in self.db_cluster.nodes:
node.run_nodetool(sub_cmd="repair -pr")
node: BaseNode = self.db_cluster.nodes[0]
self.log.info('force offstrategy on %s before running running a major compaction on node %s', self.keyspace,
node.name)
node.remoter.run(
"curl -X POST --header 'Content-Type: application/json' --header 'Accept: application/json'"
f" http://127.0.0.1:10000/storage_service/keyspace_offstrategy_compaction/{self.keyspace}")
self.log.info('Run a major compaction on node %s', node.name)
node.run_nodetool("compact")
self.wait_no_compactions_running()
return self._count_sstables_and_partitions(node=node)
def test_count_sstables_after_major_compaction(self):
"""
This test run the original test_custom_time first.
Waits a duration of TTL + scan + gc_grace_seconds.
Assume TTL=36 minutes, scan=8 minutes, gc_grace_seconds=4 minutes: Total sums to: 48 minutes.
Run a repair on nodes.
Select a node and run a major compaction.
Wait for node's compactions to finish.
count and print number of sstables, partitions and sizes.
"""
# Run the stress_cmd and wait for complete.
self.test_custom_time()
# Wait for all data to be expired.
# Should wait a total of 48 minutes.
# splitting to 30 minutes wait before stopping nemesis.
# Then an additional 18 minutes wait to complete a total of 48 minutes.
self.log.info('Wait TTL + scan interval + gc_grace_seconds')
wait_for_expired_items = 30 * 60
time.sleep(wait_for_expired_items)
self.stop_nemesis(self.db_cluster)
wait_for_expired_items = 18 * 60
time.sleep(wait_for_expired_items)
# Run a repair + a major compaction and count results.
self.count_sstables_and_partitions_after_major_compaction()
def test_custom_time_repeat_stress_cmd(self):
"""
This test run the original test_custom_time first.
Then after finished it re-execute the same stress_cmd again.
An example for a relevant scenario for that is for Alternator TTL that requires overwriting existing TTL data
after a while, where some items are already expired and/or scanned.
"""
# Run the stress_cmd a first time
self.test_custom_time()
# Rerun the stress_cmd a second time
stress_cmd = self.params.get('stress_cmd')
if stress_cmd:
stress_queue = []
params = {'stress_cmd': stress_cmd, 'round_robin': self.params.get('round_robin')}
self._run_all_stress_cmds(stress_queue, params)
for stress in stress_queue:
self.verify_stress_thread(cs_thread_pool=stress)
self.count_sstables_and_partitions_after_major_compaction()
def test_disable_enable_ttl_scan(self):
"""
1. Run pre-create schema for Alternator, disabling TTL for tables.
2. Run the original test_custom_time prepare step.
3. Enable TTL for tables.
4. Wait for TTL-scan intervals to run
5. Run a background read stress while data is being expired.
"""
# pylint: disable=too-many-locals,too-many-branches,too-many-statements
self.db_cluster.add_nemesis(nemesis=self.get_nemesis_class(),
tester_obj=self)
stress_queue = []
# prepare write workload
prepare_write_cmd = self.params.get('prepare_write_cmd')
keyspace_num = 1
self.pre_create_alternator_tables()
# Disable TTL
self.alternator.modify_alternator_ttl_spec(enabled=False, node=self.db_cluster.nodes[0])
# Run write stress
self.run_prepare_write_cmd()
stress_cmd = self.params.get('stress_cmd')
if stress_cmd:
params = {'keyspace_num': keyspace_num, 'stress_cmd': stress_cmd,
'round_robin': self.params.get('round_robin')}
self._run_all_stress_cmds(stress_queue, params)
if not prepare_write_cmd or not self.params.get('nemesis_during_prepare'):
self.db_cluster.start_nemesis()
for stress in stress_queue:
self.verify_stress_thread(cs_thread_pool=stress)
# Enable TTL
self.alternator.modify_alternator_ttl_spec(enabled=True, node=self.db_cluster.nodes[0])
stress_queue = []
# Run read stress as a background thread while TTL scans are going over existing data.
stress_read_cmd = self.params.get('stress_read_cmd')
if stress_read_cmd:
params = {'keyspace_num': keyspace_num, 'stress_cmd': stress_read_cmd}
self._run_all_stress_cmds(stress_queue, params)
for stress in stress_queue:
self.verify_stress_thread(cs_thread_pool=stress)
self.count_sstables_and_partitions_after_major_compaction()
def test_multiple_ttl(self):
"""
This test run the original test_custom_time first.
It assumes multiple TTL values to run in stress commands.
It then waits a TTL+scan-interval duration for the second longest TTL stress:
2.5 hours (second longer stress TTL) + 4 minutes (scan-interval) + 10 minutes (max estimated scan duration)
This sums up to: 2 hours and 45 minutes (=165 minutes).
Then after finished it run a table full scan.
It validates only the data of longest TTL stress (10 days) exist.
"""
# Run the stress_cmd and wait for finish
self.test_custom_time()
self.log.info('Wait second-longer-stress TTL + scan-interval + max-estimated-scan-duration')
wait_for_expired_items = 165 * 60
time.sleep(wait_for_expired_items)
stress_cmd = self.params.get('stress_cmd')
longest_ttl_stress_cmd = stress_cmd[-1]
insert_count = [param for param in longest_ttl_stress_cmd.split() if 'insertcount=' in param]
insert_count = int(insert_count[0].split('=')[1])
self.log.info('The longest-TTL-load number of items is: %s', insert_count)
self.log.info('Run a major compaction on nodes for all expected expired data tombstones to be deleted.')
for node in self.db_cluster.nodes:
node.run_nodetool("compact")
self.wait_no_compactions_running()
self.log.info('Run a table scan to count number of existing items and compare to expected')
current_partitions_num = self._count_sstables_and_partitions(node=self.db_cluster.nodes[0])
assert current_partitions_num == insert_count, f"Result: {current_partitions_num}, Expected: {insert_count}"