Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Usenet -> c.d.o.server -> Enqueue Wait Events - TM - SSX
Good day,
We have 3-node Oracle 9.2.0.5 RAC database on AIX 5.2.
We're seeing enqueue wait events in our database. Here is a Statspack output (elapsed time of about 5 minutes):
Event Waits Time (s) Ela Time ----------------------------- ------------ ----------- -------- enqueue 9,523 4,097 42.48 ksxr poll remote instances 23,000 3,512 36.41 db file sequential read 270,913 1,315 13.63 CPU time 527 5.47 global cache cr request 144,482 65 .67 -------------------------------------------------------------
Enqueue activity for DB:
-> Enqueue stats gathered prior to 9i should not be compared with 9i
data
-> ordered by Wait Time desc, Waits desc
Avg Wt Wait Eq Requests Succ Gets Failed Gets Waits Time (ms) Time (s) -- --------- ---------- ----------- ----------- ----------- ---------- TM 77,597 77,579 0 1,146 3,681.83 4,219 TX 16,795 16,793 0 31 8.32 0 CU 10,130 10,129 0 10 18.00 0 **************************************************************************
Furthermore, we have captured the running statement and its waiting
state by querying V$SESSION and V$SESSION_WAIT (and v$sql +
dba_objects):
Lo Mode OBJECT_NAME (WAIT_SECS) SQL_TEXT
-- ----- -------------- ---------------------------------------- TM 5 BWPHO_PK (3) DELETE BORROWER WHERE BORROWERKEY = :B1 TM 5 BWPHO_BORRW_FK (2) DELETE BORROWER WHERE BORROWERKEY = :B1
Lo == Lock Type
Mode == Requested Lock Mode (5 == SSX or Sub Shared Exclusive)
OJBECT_NAME == ROW_WAIT_OBJ# (obj associated with locked resource)
WAIT_SECS == seconds in wait state
SQL_TEXT == sql statement
The above output implies that the DELETE BORROWER statement is waiting on child records. We have confirmed that the referenced foreign key / primary key indexes do exist and valid.
Can anyone comment or answer on the following questions:
1) What is the difference between TM and TX lock type? 2) When would we see TM / Sub Shared Exclusive (lock mode 5)? 3) According to the Statspack output, this wait event seems significant(42% of all wait time). Is this a real wait event (one that would impact response time significantly)?
We're trying to debug this condition, and we just cannot seem to replicate it in a controlled environment.
Thanks in advance for any feedback on this.
Regards,
Kevin Lee
Optizon, Inc.
Received on Wed Jan 25 2006 - 20:26:24 CST
![]() |
![]() |