| 1 |
################################################################################ |
| 2 |
# Verify that when a member is blocked on a minority and waiting on a blocked |
| 3 |
# query, group replication will kill all transactions after a while and leave |
| 4 |
# the group enabling super read mode. |
| 5 |
# Steps: |
| 6 |
# 0) The test requires three servers: M1, M2 and M3. |
| 7 |
# 1) Create a group with 3 members, we want fourth server for async channel. |
| 8 |
# 2) Check the group_replication_unreachable_majority_timeout var boundaries |
| 9 |
# 2.1) Check min value of group_replication_unreachable_majority_timeout |
| 10 |
# 2.1) Check max value of group_replication_unreachable_majority_timeout |
| 11 |
# 3) Create a minority situation with only 1 member alive: M1 |
| 12 |
# Assert M2 and M3 are unreachable |
| 13 |
# 4) Execute a query on M1 that will block |
| 14 |
# 5) Wait a minute and assert |
| 15 |
# * member state is ERROR |
| 16 |
# * super read only mode is ON |
| 17 |
# * the query rolled back |
| 18 |
# 6) Confirm on server1 Slave IO and SQL threads are not running. |
| 19 |
# 7) Clean up. |
| 20 |
################################################################################ |
| 21 |
# Don't test this under valgrind, memory leaks will occur |
| 22 |
--source include/not_valgrind.inc |
| 23 |
--source include/big_test.inc |
| 24 |
--source include/force_restart.inc |
| 25 |
--source include/have_group_replication_plugin.inc |
| 26 |
|
| 27 |
--echo # |
| 28 |
--echo # 1. Create a group with 3 members, |
| 29 |
--echo # we want fourth server for async channel. |
| 30 |
--echo # Extract each server uuid |
| 31 |
--echo # Create a table on it. |
| 32 |
--echo # |
| 33 |
|
| 34 |
--let $rpl_server_count= 4 |
| 35 |
--source include/group_replication.inc |
| 36 |
|
| 37 |
--let $rpl_connection_name= server1 |
| 38 |
--source include/rpl_connection.inc |
| 39 |
--let $member1_uuid= query_get_value(SELECT @@SERVER_UUID, @@SERVER_UUID, 1) |
| 40 |
--let $local_address_server1= `SELECT @@GLOBAL.group_replication_local_address` |
| 41 |
|
| 42 |
--let $rpl_connection_name= server2 |
| 43 |
--source include/rpl_connection.inc |
| 44 |
--let $member2_uuid= query_get_value(SELECT @@SERVER_UUID, @@SERVER_UUID, 1) |
| 45 |
|
| 46 |
--let $rpl_connection_name= server3 |
| 47 |
--source include/rpl_connection.inc |
| 48 |
--let $member3_uuid= query_get_value(SELECT @@SERVER_UUID, @@SERVER_UUID, 1) |
| 49 |
|
| 50 |
--let $rpl_connection_name= server4 |
| 51 |
--source include/rpl_connection.inc |
| 52 |
--source include/stop_group_replication.inc |
| 53 |
|
| 54 |
--let $rpl_connection_name= server1 |
| 55 |
--source include/rpl_connection.inc |
| 56 |
--replace_result $SERVER_MYPORT_4 SERVER_4_PORT |
| 57 |
--eval CHANGE REPLICATION SOURCE TO SOURCE_HOST='127.0.0.1', SOURCE_USER='root', SOURCE_PORT=$SERVER_MYPORT_4 for channel 'ch4_1' |
| 58 |
|
| 59 |
--let $rpl_channel_name= 'ch4_1' |
| 60 |
--source include/start_slave.inc |
| 61 |
--let $rpl_channel_name= |
| 62 |
|
| 63 |
CREATE TABLE t1 (c1 INT NOT NULL PRIMARY KEY) ENGINE=InnoDB; |
| 64 |
--source include/rpl_sync.inc |
| 65 |
|
| 66 |
--echo # |
| 67 |
--echo # 2. Check group_replication_unreachable_majority_timeout boundaries |
| 68 |
--echo # |
| 69 |
|
| 70 |
--let $rpl_connection_name= server1 |
| 71 |
--source include/rpl_connection.inc |
| 72 |
|
| 73 |
--let $saved_unr_timeout= `SELECT @@GLOBAL.group_replication_unreachable_majority_timeout` |
| 74 |
|
| 75 |
--echo # 2.1. Set group_replication_unreachable_majority_timeout below min value |
| 76 |
SET @@GLOBAL.group_replication_unreachable_majority_timeout= -1; |
| 77 |
|
| 78 |
--echo # 2.1. Check min value of group_replication_unreachable_majority_timeout |
| 79 |
--let $assert_text= Minimum group_replication_unreachable_majority_timeout is 0 |
| 80 |
--let $assert_cond= "[SELECT @@GLOBAL.group_replication_unreachable_majority_timeout]" = 0 |
| 81 |
--source include/assert.inc |
| 82 |
|
| 83 |
--echo # 2.2. Set group_replication_unreachable_majority_timeout above max value. |
| 84 |
SET @@GLOBAL.group_replication_unreachable_majority_timeout= 31536001; |
| 85 |
|
| 86 |
--echo # 2.2. Check max value of group_replication_unreachable_majority_timeout. |
| 87 |
--let $assert_text= Maximum group_replication_unreachable_majority_timeout is 31536000 |
| 88 |
--let $assert_cond= "[SELECT @@GLOBAL.group_replication_unreachable_majority_timeout]" = 31536000 |
| 89 |
|
| 90 |
# Set a low value for the test |
| 91 |
SET @@GLOBAL.group_replication_unreachable_majority_timeout= 11; |
| 92 |
|
| 93 |
--echo # |
| 94 |
--echo # 3. Crash server 2 and server 3 |
| 95 |
--echo # Check they are marked as unreachable |
| 96 |
--echo # |
| 97 |
# We do kill the servers, using shutdown_server 0, and then MTR will |
| 98 |
# follow the expect file and restart the server, but we do not wait |
| 99 |
# for the server restart. |
| 100 |
# Since the servers loose its GCS configuration them will not rejoin |
| 101 |
# the group. |
| 102 |
|
| 103 |
--let $rpl_connection_name= server2 |
| 104 |
--source include/rpl_connection.inc |
| 105 |
--exec echo "restart" > $MYSQLTEST_VARDIR/tmp/mysqld.2.expect |
| 106 |
--shutdown_server 0 |
| 107 |
--source include/wait_until_disconnected.inc |
| 108 |
|
| 109 |
--let $rpl_connection_name= server3 |
| 110 |
--source include/rpl_connection.inc |
| 111 |
--exec echo "restart" > $MYSQLTEST_VARDIR/tmp/mysqld.3.expect |
| 112 |
--shutdown_server 0 |
| 113 |
--source include/wait_until_disconnected.inc |
| 114 |
|
| 115 |
--let $rpl_connection_name= server1 |
| 116 |
--source include/rpl_connection.inc |
| 117 |
|
| 118 |
--let $group_replication_member_state= UNREACHABLE |
| 119 |
--let $group_replication_member_id= $member2_uuid |
| 120 |
--source include/gr_wait_for_member_state.inc |
| 121 |
|
| 122 |
--let $group_replication_member_state= UNREACHABLE |
| 123 |
--let $group_replication_member_id= $member3_uuid |
| 124 |
--source include/gr_wait_for_member_state.inc |
| 125 |
|
| 126 |
--let $group_replication_member_id= $member1_uuid |
| 127 |
|
| 128 |
--echo # |
| 129 |
--echo # 4. Execute a query that will block |
| 130 |
--echo # |
| 131 |
|
| 132 |
--send INSERT INTO t1 VALUES (2) |
| 133 |
|
| 134 |
--echo # |
| 135 |
--echo # 5. After some seconds the member should go into error mode. |
| 136 |
--echo # It should enable super read only mode and kill pending transactions. |
| 137 |
--echo # |
| 138 |
|
| 139 |
--let $rpl_connection_name= server_1 |
| 140 |
--source include/rpl_connection.inc |
| 141 |
|
| 142 |
# verify thread THD_group_partition_handler has started |
| 143 |
--let $wait_condition= SELECT COUNT(*)=1 FROM performance_schema.threads WHERE NAME = "thread/group_rpl/THD_group_partition_handler" |
| 144 |
--source include/wait_condition.inc |
| 145 |
|
| 146 |
--let $group_replication_member_state= ERROR |
| 147 |
--source include/gr_wait_for_member_state.inc |
| 148 |
|
| 149 |
--let $rpl_connection_name= server1 |
| 150 |
--source include/rpl_connection.inc |
| 151 |
|
| 152 |
--error ER_RUN_HOOK_ERROR |
| 153 |
--reap |
| 154 |
|
| 155 |
--echo # The server's read_only mode should be enabled |
| 156 |
--let $assert_text= The super_read_only mode should be 1. |
| 157 |
--let $assert_cond= "[SELECT @@GLOBAL.super_read_only]" = 1; |
| 158 |
--source include/assert.inc |
| 159 |
|
| 160 |
|
| 161 |
--echo |
| 162 |
--echo # 6. Confirm on server1 Slave IO and SQL threads are not running. |
| 163 |
|
| 164 |
--let $slave_param= Slave_IO_Running |
| 165 |
--let $slave_param_value= No |
| 166 |
--source include/check_slave_param.inc |
| 167 |
|
| 168 |
--let $slave_param= Slave_SQL_Running |
| 169 |
--let $slave_param_value= No |
| 170 |
--source include/check_slave_param.inc |
| 171 |
|
| 172 |
|
| 173 |
--echo # |
| 174 |
--echo # 7. Clean up. |
| 175 |
--echo # |
| 176 |
|
| 177 |
--source include/stop_group_replication.inc |
| 178 |
|
| 179 |
--let $rpl_connection_name= server1 |
| 180 |
--source include/rpl_connection.inc |
| 181 |
|
| 182 |
DROP TABLE t1; |
| 183 |
|
| 184 |
--disable_query_log |
| 185 |
--eval SET @@GLOBAL.group_replication_unreachable_majority_timeout= $saved_unr_timeout |
| 186 |
--enable_query_log |
| 187 |
|
| 188 |
set session sql_log_bin=0; |
| 189 |
call mtr.add_suppression("The member lost contact with a majority of the members in the group. Until the network is restored.*"); |
| 190 |
call mtr.add_suppression("'This member could not reach a majority of the members for more than.*"); |
| 191 |
call mtr.add_suppression("The server was automatically set into read only mode after an error was detected."); |
| 192 |
call mtr.add_suppression("Due to a plugin error, some transactions were unable to be certified and will now rollback."); |
| 193 |
call mtr.add_suppression("Error while waiting for conflict detection procedure to finish on session.*"); |
| 194 |
call mtr.add_suppression("Run function 'before_commit' in plugin 'group_replication' failed"); |
| 195 |
call mtr.add_suppression("Skipping leave operation: concurrent attempt to leave the group is on-going."); |
| 196 |
call mtr.add_suppression("Timeout while waiting for the group communication engine to exit!"); |
| 197 |
call mtr.add_suppression("The member has failed to gracefully leave the group."); |
| 198 |
call mtr.add_suppression("read failed"); |
| 199 |
set session sql_log_bin=1; |
| 200 |
|
| 201 |
--let $rpl_connection_name= server_2 |
| 202 |
--source include/rpl_connection.inc |
| 203 |
|
| 204 |
--let $rpl_server_number= 2 |
| 205 |
--source include/rpl_reconnect.inc |
| 206 |
|
| 207 |
DROP TABLE t1; |
| 208 |
|
| 209 |
--let $rpl_connection_name= server_3 |
| 210 |
--source include/rpl_connection.inc |
| 211 |
|
| 212 |
--let $rpl_server_number= 3 |
| 213 |
--source include/rpl_reconnect.inc |
| 214 |
|
| 215 |
DROP TABLE t1; |
| 216 |
|
| 217 |
--source include/group_replication_end.inc |