. ./init.sh # Needs to apply lots of relay log files # Relay log files switched many times while master binary log was not switched because max_relay_log_size is so small. In this case, each relay log does not contain a Rotate Event. This test case is to check recover works without rotate event. mysql $S1 -e "set global max_relay_log_size=8192" mysql $S2 -e "set global max_relay_log_size=8192" mysql $S3 -e "set global max_relay_log_size=8192" mysql $S4 -e "set global max_relay_log_size=8192" mysql $S1 test -e "set global relay_log_purge=0" mysql $S2 test -e "set global relay_log_purge=0" mysql $S3 test -e "set global relay_log_purge=0" mysql $S4 test -e "set global relay_log_purge=0" perl insert.pl $MP $MYSQL_USER $MYSQL_PWD 2 1000 0 sleep 2 mysql $S1 test -e "stop slave io_thread" perl insert.pl $MP $MYSQL_USER $MYSQL_PWD 1001 3000 0 sleep 1 ./kill_m.sh ./run.sh fail_if_nonzero $0 $? mysql $S1 test -e "insert into t1 values(10000003, 300, 'bbbaaaaaaa');" ./check $0 3001