Debian Codegraph @

dirsrvtests/tests/suites/replication/cleanallruv_abort_restart_test.py

raw
1 # --- BEGIN COPYRIGHT BLOCK ---
2 # Copyright (C) 2022 Red Hat, Inc.
3 # All rights reserved.
4 #
5 # License: GPL (version 3 or any later version).
6 # See LICENSE for details.
7 # --- END COPYRIGHT BLOCK ---
8 #
9 import logging
10 import pytest
11 import os
12 import time
13 from lib389._constants import DEFAULT_SUFFIX
14 from test389.topologies import topology_m4
15 from lib389.tasks import CleanAllRUVTask
16 from lib389.replica import ReplicationManager
17
18 log = logging.getLogger(__name__)
19
20
21 def remove_supplier4_agmts(msg, topology_m4):
22 """Remove all the repl agmts to supplier4. """
23
24 log.info('%s: remove all the agreements to supplier 4...' % msg)
25 repl = ReplicationManager(DEFAULT_SUFFIX)
26 # This will delete m4 from the topo *and* remove all incoming agreements
27 # to m4.
28 repl.remove_supplier(topology_m4.ms["supplier4"],
29 [topology_m4.ms["supplier1"], topology_m4.ms["supplier2"], topology_m4.ms["supplier3"]])
30
31 def task_done(topology_m4, task_dn, timeout=60):
32 """Check if the task is complete"""
33
34 attrlist = ['nsTaskLog', 'nsTaskStatus', 'nsTaskExitCode',
35 'nsTaskCurrentItem', 'nsTaskTotalItems']
36 done = False
37 count = 0
38
39 while not done and count < timeout:
40 try:
41 entry = topology_m4.ms["supplier1"].getEntry(task_dn, attrlist=attrlist)
42 if entry is not None:
43 if entry.hasAttr('nsTaskExitCode'):
44 done = True
45 break
46 else:
47 done = True
48 break
49 except ldap.NO_SUCH_OBJECT:
50 done = True
51 break
52 except ldap.LDAPError:
53 break
54 time.sleep(1)
55 count += 1
56
57 return done
58
59
60 @pytest.mark.flaky(max_runs=2, min_passes=1)
61 def test_abort_restart(topology_m4):
62 """Test the abort task can handle a restart, and then resume
63
64 :id: b66e33d4-fe85-4e1c-b882-75da80f70ab3
65 :setup: Replication setup with four suppliers
66 :steps:
67 1. Disable replication on supplier 4
68 2. Remove agreements to supplier 4 from other suppliers
69 3. Stop supplier 3
70 4. Run a cleanallruv task on supplier 1
71 5. Run a cleanallruv abort task on supplier 1
72 6. Restart supplier 1
73 7. Make sure that no crash happened
74 8. Start supplier 3
75 9. Check supplier 1 does not have the clean task running
76 10. Check that errors log doesn't have 'Aborting abort task' message
77 :expectedresults:
78 1. Replication on supplier 4 should be disabled
79 2. Agreements to supplier 4 should be removed
80 3. Supplier 3 should be stopped
81 4. Operation should be successful
82 5. Operation should be successful
83 6. Supplier 1 should be restarted
84 7. No crash should happened
85 8. Supplier 3 should be started
86 9. Check supplier 1 shouldn't have the clean task running
87 10. Errors log shouldn't have 'Aborting abort task' message
88 """
89
90 log.info('Running test_abort_restart...')
91 # Remove the agreements from the other suppliers that point to supplier 4
92 repl = ReplicationManager(DEFAULT_SUFFIX)
93 m4rid = repl.get_rid(topology_m4.ms["supplier4"])
94 remove_supplier4_agmts("test_abort", topology_m4)
95
96 # Stop supplier 3
97 log.info('test_abort_restart: stop supplier 3 to freeze the cleanAllRUV task...')
98 topology_m4.ms["supplier3"].stop()
99
100 # Run the task
101 log.info('test_abort_restart: add the cleanAllRUV task...')
102 cruv_task = CleanAllRUVTask(topology_m4.ms["supplier1"])
103 cruv_task.create(properties={
104 'replica-id': m4rid,
105 'replica-base-dn': DEFAULT_SUFFIX,
106 'replica-force-cleaning': 'no',
107 'replica-certify-all': 'yes'
108 })
109 # Wait a bit
110 time.sleep(2)
111
112 # Abort the task
113 cruv_task.abort(certify=True)
114
115 # Check supplier 1 does not have the clean task running
116 log.info('test_abort_abort: check supplier 1 no longer has a cleanAllRUV task...')
117 if not task_done(topology_m4, cruv_task.dn):
118 log.fatal('test_abort_restart: CleanAllRUV task was not aborted')
119 assert False
120
121 # Now restart supplier 1, and make sure the abort process completes
122 topology_m4.ms["supplier1"].restart()
123 if topology_m4.ms["supplier1"].detectDisorderlyShutdown():
124 log.fatal('test_abort_restart: Supplier 1 previously crashed!')
125 assert False
126
127 # Start supplier 3
128 topology_m4.ms["supplier3"].start()
129
130 # Need to wait 5 seconds before server processes any leftover tasks
131 time.sleep(6)
132
133 # Check supplier 1 tried to run abort task. We expect the abort task to be aborted.
134 if not topology_m4.ms["supplier1"].searchErrorsLog('Aborting abort task'):
135 log.fatal('test_abort_restart: Abort task did not restart')
136 assert False
137
138 log.info('test_abort_restart PASSED')
139
140
141 if __name__ == '__main__':
142 # Run isolated
143 # -s for DEBUG mode
144 CURRENT_FILE = os.path.realpath(__file__)
145 pytest.main(["-s", CURRENT_FILE])
146