| 1 |
################################################################################ |
| 2 |
# Validate that when a server joins a group and is not able to find |
| 3 |
# a suitable donor it errors out and does not block on join. |
| 4 |
# |
| 5 |
# Test: |
| 6 |
# 0. The test requires three servers: M1, M2 and M3. |
| 7 |
# 1. Create a group with three members and schedule a periodic event to create |
| 8 |
# some load in the group on M1. |
| 9 |
# 2. Add suppressions. |
| 10 |
# 3. Block applier channel on M1 so that joinining members that will recover |
| 11 |
# from it will be blocked on recovery waiting for the View_change_log_event. |
| 12 |
# 4. Stop and start M2 and M3. M2 and M3 will be in recovery. |
| 13 |
# 5. Stop and start M1. M1 will be in recovery since all existent group members |
| 14 |
# are RECOVERING. |
| 15 |
# 6. Clean up. |
| 16 |
################################################################################ |
| 17 |
--source include/big_test.inc |
| 18 |
--source include/have_debug_sync.inc |
| 19 |
--source include/set_privilege_checks_user_as_system_user.inc |
| 20 |
--source include/have_group_replication_plugin.inc |
| 21 |
|
| 22 |
|
| 23 |
--echo |
| 24 |
--echo ############################################################ |
| 25 |
--echo # 1. Create a group with 1 member and schedule a periodic |
| 26 |
--echo # event to create some load in the group. |
| 27 |
--let $rpl_server_count= 3 |
| 28 |
--source include/group_replication.inc |
| 29 |
|
| 30 |
--let $rpl_connection_name= server1 |
| 31 |
--source include/rpl_connection.inc |
| 32 |
--let $server1_uuid= `SELECT @@GLOBAL.SERVER_UUID` |
| 33 |
SET SESSION sql_log_bin= 0; |
| 34 |
call mtr.add_suppression("On shutdown there was a timeout on the Group Replication recovery module termination. Check the log for more details"); |
| 35 |
call mtr.add_suppression("Fatal error during the incremental recovery process of Group Replication. The server will leave the group."); |
| 36 |
call mtr.add_suppression("The server was automatically set into read only mode after an error was detected."); |
| 37 |
call mtr.add_suppression("Error leaving the group"); |
| 38 |
call mtr.add_suppression("All donors left. Aborting group replication incremental recovery."); |
| 39 |
call mtr.add_suppression("This member has more executed transactions *.*"); |
| 40 |
call mtr.add_suppression("The member contains transactions not present in the group. It is only allowed to *.*"); |
| 41 |
call mtr.add_suppression("The member is already leaving or joining a group."); |
| 42 |
call mtr.add_suppression("Skipping leave operation: member already left the group."); |
| 43 |
call mtr.add_suppression("Skipping leave operation: concurrent attempt to leave the group is on-going."); |
| 44 |
call mtr.add_suppression("Maximum number of retries when trying to connect to a donor reached. Aborting group replication incremental recovery."); |
| 45 |
call mtr.add_suppression("Transaction cannot be executed while Group Replication is stopping."); |
| 46 |
call mtr.add_suppression("Run function 'before_commit' in plugin 'group_replication' failed"); |
| 47 |
call mtr.add_suppression("While leaving the group due to a stop, shutdown or failure there was a timeout receiving a view change. This can lead to a possible inconsistent state. Check the log for more details"); |
| 48 |
call mtr.add_suppression("On shutdown there was a timeout on the Group Replication applier termination."); |
| 49 |
SET SESSION sql_log_bin= 1; |
| 50 |
|
| 51 |
CREATE TABLE t1 (a INT PRIMARY KEY AUTO_INCREMENT NOT NULL); |
| 52 |
SET @old_event_scheduler= @@global.event_scheduler; |
| 53 |
SET GLOBAL EVENT_SCHEDULER= ON; |
| 54 |
delimiter //; |
| 55 |
CREATE EVENT IF NOT EXISTS e1 |
| 56 |
ON SCHEDULE |
| 57 |
EVERY 1 SECOND |
| 58 |
STARTS CURRENT_TIMESTAMP + INTERVAL 1 SECOND |
| 59 |
ENDS CURRENT_TIMESTAMP + INTERVAL 40 SECOND |
| 60 |
ON COMPLETION PRESERVE |
| 61 |
ENABLE |
| 62 |
COMMENT 'Test with events' |
| 63 |
DO |
| 64 |
BEGIN |
| 65 |
DECLARE v INTEGER; |
| 66 |
DECLARE CONTINUE HANDLER FOR SQLEXCEPTION BEGIN END; |
| 67 |
|
| 68 |
SET v=0; |
| 69 |
|
| 70 |
WHILE v < 10 |
| 71 |
DO |
| 72 |
INSERT INTO test.t1 values (NULL); |
| 73 |
SET v = v+1; |
| 74 |
END WHILE; |
| 75 |
END// |
| 76 |
|
| 77 |
delimiter ;// |
| 78 |
|
| 79 |
|
| 80 |
--echo |
| 81 |
--echo ############################################################ |
| 82 |
--echo # 2. Add error suppressions. |
| 83 |
--let $rpl_connection_name= server1 |
| 84 |
--source include/rpl_connection.inc |
| 85 |
SET SESSION sql_log_bin = 0; |
| 86 |
call mtr.add_suppression("Replica SQL for channel 'group_replication_applier': Relay log read failure: Could not parse relay log event entry.*"); |
| 87 |
call mtr.add_suppression("The applier thread execution was aborted. Unable to process more transactions, this member will now leave the group."); |
| 88 |
call mtr.add_suppression("Fatal error during execution on the Applier process of Group Replication. The server will now leave the group."); |
| 89 |
call mtr.add_suppression("The server was automatically set into read only mode after an error was detected."); |
| 90 |
call mtr.add_suppression("Skipping leave operation: concurrent attempt to leave the group is on-going."); |
| 91 |
SET SESSION sql_log_bin = 1; |
| 92 |
|
| 93 |
--let $rpl_connection_name= server2 |
| 94 |
--source include/rpl_connection.inc |
| 95 |
SET SESSION sql_log_bin= 0; |
| 96 |
call mtr.add_suppression("On shutdown there was a timeout on the Group Replication recovery module termination. Check the log for more details"); |
| 97 |
call mtr.add_suppression("Replica SQL for channel 'group_replication_recovery': Could not execute Write_rows event on table test.t1; Duplicate entry.*"); |
| 98 |
call mtr.add_suppression("Fatal error during the incremental recovery process of Group Replication. The server will leave the group."); |
| 99 |
call mtr.add_suppression("The server was automatically set into read only mode after an error was detected."); |
| 100 |
call mtr.add_suppression("Error leaving the group"); |
| 101 |
call mtr.add_suppression("All donors left. Aborting group replication incremental recovery."); |
| 102 |
call mtr.add_suppression("The member is already leaving or joining a group."); |
| 103 |
call mtr.add_suppression("Skipping leave operation: member already left the group."); |
| 104 |
call mtr.add_suppression("Skipping leave operation: concurrent attempt to leave the group is on-going."); |
| 105 |
SET SESSION sql_log_bin= 1; |
| 106 |
|
| 107 |
--let $rpl_connection_name= server3 |
| 108 |
--source include/rpl_connection.inc |
| 109 |
SET SESSION sql_log_bin= 0; |
| 110 |
call mtr.add_suppression("On shutdown there was a timeout on the Group Replication recovery module termination. Check the log for more details"); |
| 111 |
call mtr.add_suppression("Replica SQL for channel 'group_replication_recovery': Could not execute Write_rows event on table test.t1; Duplicate entry.*"); |
| 112 |
call mtr.add_suppression("Fatal error during the incremental recovery process of Group Replication. The server will leave the group."); |
| 113 |
call mtr.add_suppression("The server was automatically set into read only mode after an error was detected."); |
| 114 |
call mtr.add_suppression("Error leaving the group"); |
| 115 |
call mtr.add_suppression("All donors left. Aborting group replication incremental recovery."); |
| 116 |
call mtr.add_suppression("The member is already leaving or joining a group."); |
| 117 |
call mtr.add_suppression("Skipping leave operation: member already left the group."); |
| 118 |
call mtr.add_suppression("Skipping leave operation: concurrent attempt to leave the group is on-going."); |
| 119 |
--disable_query_log # No need for the result file to specify suppressions. |
| 120 |
call mtr.add_suppression("While leaving the group due to a stop, shutdown or failure there was a timeout receiving a view change. This can lead to a possible inconsistent state. Check the log for more details"); |
| 121 |
--enable_query_log |
| 122 |
SET SESSION sql_log_bin= 1; |
| 123 |
|
| 124 |
|
| 125 |
--echo |
| 126 |
--echo ############################################################ |
| 127 |
--echo # 3. Block applier channel on member 1 so that joinining |
| 128 |
--echo # members that will recover from it will be blocked on |
| 129 |
--echo # recovery waiting for the View_change_log_event on which |
| 130 |
--echo # them joined. |
| 131 |
--let $rpl_connection_name= server1 |
| 132 |
--source include/rpl_connection.inc |
| 133 |
SET @@GLOBAL.DEBUG='+d,block_applier_updates'; |
| 134 |
|
| 135 |
|
| 136 |
--echo |
| 137 |
--echo ############################################################ |
| 138 |
--echo # 4. Stop and start members 2 and 3 without waiting for |
| 139 |
--echo # member status. |
| 140 |
--let $rpl_connection_name= server2 |
| 141 |
--source include/rpl_connection.inc |
| 142 |
STOP GROUP_REPLICATION; |
| 143 |
--let $rpl_connection_name= server3 |
| 144 |
--source include/rpl_connection.inc |
| 145 |
STOP GROUP_REPLICATION; |
| 146 |
|
| 147 |
--let $rpl_connection_name= server2 |
| 148 |
--source include/rpl_connection.inc |
| 149 |
--source include/start_group_replication_command.inc |
| 150 |
--let $rpl_connection_name= server3 |
| 151 |
--source include/rpl_connection.inc |
| 152 |
--source include/start_group_replication_command.inc |
| 153 |
|
| 154 |
# Member 2 and 3 will be on recovery |
| 155 |
--let $rpl_connection_name= server2 |
| 156 |
--source include/rpl_connection.inc |
| 157 |
--let $group_replication_member_state= RECOVERING |
| 158 |
--source include/gr_wait_for_member_state.inc |
| 159 |
|
| 160 |
--let $rpl_connection_name= server3 |
| 161 |
--source include/rpl_connection.inc |
| 162 |
--let $group_replication_member_state= RECOVERING |
| 163 |
--source include/gr_wait_for_member_state.inc |
| 164 |
|
| 165 |
|
| 166 |
--echo |
| 167 |
--echo ############################################################ |
| 168 |
--echo # 5. Stop and start member 1. |
| 169 |
--let $rpl_connection_name= server1 |
| 170 |
--source include/rpl_connection.inc |
| 171 |
SET @group_replication_recovery_reconnect_interval_save= @@GLOBAL.group_replication_recovery_reconnect_interval; |
| 172 |
SET GLOBAL group_replication_recovery_reconnect_interval= 1; |
| 173 |
SET @group_replication_components_stop_timeout_save = @@GLOBAL.group_replication_components_stop_timeout; |
| 174 |
SET GLOBAL group_replication_components_stop_timeout= 2; |
| 175 |
|
| 176 |
SET DEBUG_SYNC = "now WAIT_FOR applier_read_blocked"; |
| 177 |
SET @@GLOBAL.DEBUG='-d,block_applier_updates'; |
| 178 |
SET @@GLOBAL.DEBUG='+d,force_sql_thread_error'; |
| 179 |
SET DEBUG_SYNC = "now SIGNAL resume_applier_read"; |
| 180 |
--let $group_replication_member_state= ERROR |
| 181 |
--source include/gr_wait_for_member_state.inc |
| 182 |
SET @@GLOBAL.DEBUG='-d,force_sql_thread_error'; |
| 183 |
SET DEBUG_SYNC= 'RESET'; |
| 184 |
--error 0, ER_GROUP_REPLICATION_STOP_APPLIER_THREAD_TIMEOUT |
| 185 |
STOP GROUP_REPLICATION; |
| 186 |
--source include/assert_and_disable_read_only.inc |
| 187 |
|
| 188 |
# Remove all M1 data and force a RESET MASTER to allow server1 to |
| 189 |
# join, despite intentionally has more transactions than the group. |
| 190 |
ALTER EVENT e1 DISABLE; |
| 191 |
--let $wait_timeout= 300 |
| 192 |
--let $wait_condition= SELECT status="DISABLED" FROM information_schema.events WHERE event_schema='test' AND event_name="e1" |
| 193 |
--source include/wait_condition_or_abort.inc |
| 194 |
SET GLOBAL EVENT_SCHEDULER= @old_event_scheduler; |
| 195 |
|
| 196 |
DROP EVENT e1; |
| 197 |
DROP TABLE t1; |
| 198 |
RESET MASTER; |
| 199 |
|
| 200 |
--source include/start_group_replication_command.inc |
| 201 |
|
| 202 |
# Member 1 will be on recovery, since all existent group members |
| 203 |
# are RECOVERING. |
| 204 |
--let $group_replication_member_state= RECOVERING |
| 205 |
--source include/gr_wait_for_member_state.inc |
| 206 |
|
| 207 |
|
| 208 |
--echo |
| 209 |
--echo ############################################################ |
| 210 |
--echo # 6. Clean up. |
| 211 |
--let $rpl_connection_name= server1 |
| 212 |
--source include/rpl_connection.inc |
| 213 |
--source include/stop_group_replication.inc |
| 214 |
SET GLOBAL group_replication_recovery_reconnect_interval= @group_replication_recovery_reconnect_interval_save; |
| 215 |
SET GLOBAl group_replication_components_stop_timeout = @group_replication_components_stop_timeout_save; |
| 216 |
|
| 217 |
--let $rpl_connection_name= server2 |
| 218 |
--source include/rpl_connection.inc |
| 219 |
--source include/stop_group_replication.inc |
| 220 |
--disable_warnings |
| 221 |
DROP EVENT IF EXISTS e1; |
| 222 |
DROP TABLE IF EXISTS t1; |
| 223 |
--enable_warnings |
| 224 |
|
| 225 |
--let $rpl_connection_name= server3 |
| 226 |
--source include/rpl_connection.inc |
| 227 |
--source include/stop_group_replication.inc |
| 228 |
--disable_warnings |
| 229 |
DROP EVENT IF EXISTS e1; |
| 230 |
DROP TABLE IF EXISTS t1; |
| 231 |
--enable_warnings |
| 232 |
|
| 233 |
--source include/group_replication_end.inc |