Current Path : /home/usr.opt/mysql57/mysql-test/suite/perfschema/t/ |
FreeBSD hs32.drive.ne.jp 9.1-RELEASE FreeBSD 9.1-RELEASE #1: Wed Jan 14 12:18:08 JST 2015 root@hs32.drive.ne.jp:/sys/amd64/compile/hs32 amd64 |
Current File : //home/usr.opt/mysql57/mysql-test/suite/perfschema/t/threads_mysql.test |
# Tests for special content of performance_schema.threads # # Show MySQL server related content in performance_schema.threads --source include/not_windows.inc --source include/not_embedded.inc --source include/have_perfschema.inc # Ensure that the event scheduler (started via threads_mysql-master.opt) # is really running. --source include/running_event_scheduler.inc SET GLOBAL event_scheduler = OFF; --source include/no_running_event_scheduler.inc # threads are removed from: # - information_schema.processlist # - performance_schema.threads # at different times, so we may have to wait a little more # for the event_scheduler to shutdown # let $wait_timeout= 1; let $wait_condition= SELECT COUNT(*) = 0 FROM performance_schema.threads WHERE name like 'thread/sql/event%'; --source include/wait_condition.inc --vertical_results # Show all "system" threads except the event scheduler # --------------------------------------------------- # 0. The values within the columns "thread_id" and "processlist_id" depend # on server history. Therefore they are unstable and need to be omitted. # 1. The columns "time" and "state" are omitted because it is thinkable that # they contain slightly unstable though correct values. # 2. The unification of the column "parent_thread_id" is in the moment most # probably not required because I observed all time that the parent is # "thread/sql/main" with the thread_id = 1. # But there might be more kinds of parents with most probably unstable # "thread_id" values in future. # 3. In case the test is started with the option "--ps-protocol" we will # find a correcet row for our current thread but the content will differ. # Therefore we have to disable this protocol for the next statement. --disable_ps_protocol SELECT name, type, processlist_user, processlist_host, processlist_db, processlist_command, processlist_info, IF(parent_thread_id IS NULL, parent_thread_id, 'unified parent_thread_id') AS unified_parent_thread_id, role, instrumented FROM performance_schema.threads WHERE name LIKE 'thread/sql%' ORDER BY name; --enable_ps_protocol CREATE TEMPORARY TABLE t1 AS SELECT thread_id FROM performance_schema.threads WHERE name LIKE 'thread/sql%'; SET GLOBAL event_scheduler = ON; --source include/running_event_scheduler.inc # Show entries belonging to the just started event scheduler SELECT name, type, processlist_user, processlist_host, processlist_db, processlist_command, processlist_info, IF(parent_thread_id IS NULL, parent_thread_id, 'unified parent_thread_id') AS unified_parent_thread_id, role, instrumented FROM performance_schema.threads WHERE name LIKE 'thread/sql%' AND thread_id NOT IN (SELECT thread_id FROM t1) ORDER BY name; TRUNCATE t1; INSERT INTO t1 SELECT thread_id FROM performance_schema.threads WHERE name LIKE 'thread/sql%'; SELECT COUNT(*) INTO @aux FROM t1; # Attention: # Just waiting for some new thread showing up is not sufficient because # because the successing SELECT showing the thread might catch this thread # in a very early and short phase. # process_info si quite often # CREATE PROCEDURE `t_ps_event`() SQL SECURITY INVOKER SELECT SLEEP(3) # let $wait_condition= SELECT COUNT(*) = 1 FROM information_schema.processlist WHERE info = 'SELECT SLEEP(3)'; --disable_warnings DROP EVENT IF EXISTS t_ps_event; --enable_warnings CREATE EVENT t_ps_event ON SCHEDULE AT CURRENT_TIMESTAMP + INTERVAL 1 SECOND DO SELECT SLEEP(3); # Wait till one more thread comes up. # This must be the event worker thread. let $wait_timeout= 20; --source include/wait_condition.inc # Show entries belonging to the just started event worker SELECT name, type, processlist_user, processlist_host, processlist_db, processlist_command, processlist_info, IF(parent_thread_id IS NULL, parent_thread_id, 'unified parent_thread_id') AS unified_parent_thread_id, role, instrumented FROM performance_schema.threads WHERE name LIKE 'thread/sql%' AND thread_id NOT IN (SELECT thread_id FROM t1) ORDER BY name; # Show parent - child relations between "system" threads # ------------------------------------------------------ --horizontal_results SELECT t2.name AS parent_thread_name, t1.name AS child_thread_name FROM performance_schema.threads t1 INNER JOIN performance_schema.threads t2 ON t1.parent_thread_id = t2.thread_id WHERE t1.name LIKE 'thread/sql%' AND t1.parent_thread_id IS NOT NULL ORDER BY parent_thread_name, child_thread_name; # Cleanup # Wait till the event worker disappeared --source include/no_running_events.inc