wait for a undo record等待事件

今天在做AWR分析时,碰到了个新的等待事件:wait for a undo record

Top 5 Timed Events Avg %Total
~~~~~~~~~~~~~~~~~~ wait Call
Event Waits Time (s) (ms) Time Wait Class
------------------------------ ------------ ----------- ------ ------ ----------
wait for a undo record 54,998 3,557 65 43.0 Other
CPU time 1,687 20.4
db file sequential read 187,848 1,632 9 19.7 User I/O
wait for stopper event to be i 4,027 278 69 3.4 Other
read by other session 47,621 276 6 3.3 User I/O
-------------------------------------------------------------

查看了下当前数据库的等待事件,有大量的的Txn Recovery Start,
看样子是kill的大事务,数据库正在恢复

SQL> select sid, event, wait_class from v$session_wait s where s.EVENT not in ('SQL*Net message from client', 'rdbms ipc message');

SID EVENT WAIT_CLASS
---- ---------------------------------------------------------------- ----------------------------------------------------------------
2070 PX Deq: Txn Recovery Start Idle
2114 PX Deq: Txn Recovery Start Idle
2168 PX Deq: Txn Recovery Start Idle
2239 SQL*Net message to client Network
2249 PX Deq: Txn Recovery Start Idle
2264 jobq slave wait Idle
2283 wait for a undo record Other
2306 PX Deq: Txn Recovery Start Idle
2383 jobq slave wait Idle
2414 PX Deq: Txn Recovery Start Idle
2455 PX Deq: Txn Recovery Start Idle
2474 jobq slave wait Idle
2475 db file sequential read User I/O
2492 jobq slave wait Idle
2499 PX Deq: Txn Recovery Start Idle
2512 PX Deq: Txn Recovery Start Idle
2550 PX Deq: Txn Recovery Start Idle
2556 PX Deq: Txn Recovery Start Idle
2591 PX Deq: Txn Recovery Start Idle
2629 PX Deq: Txn Recovery Start Idle
2664 PX Deq: Txn Recovery Start Idle
2700 db file sequential read User I/O
2712 PX Deq: Txn Recovery Start Idle
2726 jobq slave wait Idle
2727 Streams AQ: waiting for time management or cleanup tasks Idle
2735 Streams AQ: qmn coordinator idle wait Idle
2736 Streams AQ: qmn slave idle wait Idle
2738 jobq slave wait Idle
2748 wait for stopper event to be increased Other

29 rows selected

查看v$fast_start_servers:

SQL> select pid, state, undoblocksdone from v$fast_start_servers;

PID STATE UNDOBLOCKSDONE
---------- ----------- --------------
157 RECOVERING 3123
322 RECOVERING 0
335 RECOVERING 0
455 RECOVERING 0
458 RECOVERING 0
460 RECOVERING 0
461 RECOVERING 0
463 RECOVERING 0
467 RECOVERING 0
468 RECOVERING 0
471 RECOVERING 0
472 RECOVERING 0
477 RECOVERING 0
480 RECOVERING 0
481 RECOVERING 0
484 RECOVERING 0

16 rows selected

SQL> select pid, state, undoblocksdone from v$fast_start_servers;

PID STATE UNDOBLOCKSDONE
---------- ----------- --------------
157 RECOVERING 3458
322 RECOVERING 0
335 RECOVERING 0
455 RECOVERING 0
458 RECOVERING 0
460 RECOVERING 0
461 RECOVERING 0
463 RECOVERING 0
467 RECOVERING 0
468 RECOVERING 0
471 RECOVERING 0
472 RECOVERING 0
477 RECOVERING 0
480 RECOVERING 0
481 RECOVERING 0
484 RECOVERING 0

16 rows selected


当数据库恢复结束,wait for a undo record等待事件消失
from:http://blog.chinaunix.net/u2/60220/showart_1950669.html

你可能感兴趣的:(wait)