Latches in top 5 wait events

Hi,
I have Oracle 8.1.7.4 running on solaris 8. Recently I bounced the DB and after restart the performance is very slow.
Most of the dictionary related queries take very long.
Top 3 wait events in statspack report are as below:
Top 3 Wait Events
~~~~~~~~~~~~~~~~~ Wait % Total
Event Waits Time (cs) Wt Time
latch free 58,325,603 128,928,423 81.29
sort segment request 200,342 20,671,460 13.03
enqueue 20,465 6,145,889 3.88
Can anyone provide me with some thoughts on it.
Thanks
Ramanbir

Optimizer is still the same CHOOSE and no stats were gathered on sys schema.
However i can see following in v$latch.
SELECT NAME, GETS, SLEEPS, round(SLEEPS*100/SUM(SLEEPS) OVER(),2) SLEEP_PCT, round(SLEEPS*100/GETS,2) SLEEP_RATE
               FROM V$LATCH WHERE GETS>0 ORDER BY SLEEPS DESC;
NAME     GETS     SLEEPS     SLEEP_PCT     SLEEP_RATE
cache buffers chains     161393126526     113344597     62.66     0.07
cache buffer handles     617499052     67425052     37.27     10.92
SGA is as below
Total System Global Area 1906311328 bytes
Fixed Size 73888 bytes
Variable Size 273317888 bytes
Database Buffers 1631854592 bytes
Redo Buffers 1064960 bytes
Raman

Similar Messages

Oracle RAC 9i LMD library cache lock top wait event

We are experiencing the library cache lock as our top wait event. Even thought the box is currently idle, The Global Enqueue Service Daemon (LMD) is taking up CPU cycles. The background process is also logging to trace "skgxpdocon: warning outstanding accept handle count has reached new high water mark 245000".
Any help would be appreciated.
Thanks

There is a new patch for this - check out p4673610 on metalink. We have also experience the problem in 9.2.0.8.

Top Wait Events Query is needed

Hi,
I hope I'm asking this question in right place.
I need a script and its output should give me the top 5 wait events in last 1 hour for an instance.

986330 wrote:
Hi,
I hope I'm asking this question in right place.
I need a script and its output should give me the top 5 wait events in last 1 hour for an instance.
which Top 5? Top number of Waits? Top Total time Waited? Top Avg Wait Time?
why don't you just run AWR report?

Top wait events in awr

hi,
We are using 11.2.0.3.0 on solaris 10 facing slow performance, following are the Wait Events in AWR report, need assistance to overcome it. Also if any specific document to analyze AWR report and to pin point the performance bottleneck.
Foreground Wait Events
Avg
%Time Total Wait
wait
Waits   % DB
Event
Waits -outs   Time (s)
(ms)
/txn   time
direct path read
308,729
0
21,191
69
58.0   39.5
db file sequential read
208,754
0
3,742
18
39.2
7.0
cursor: pin S
19,541,899
0
2,561
0 3,668.5
4.8
Background Wait Events
Avg
%Time Total Wait
wait
Waits   % bg
Event
Waits -outs   Time (s)
(ms)
/txn   time
log file parallel write
26,479
0
942
36
5.0   40.3
db file parallel write
216,823
0
809
4
40.7   34.6
control file sequential re
11,673
0
56
5
2.2
2.4
control file parallel writ
6,280
0
35
6
1.2
1.5
direct path read
534
0
26
49
0.1
1.1

You need to identify if you are excessively running Parallel Query -- too many queries being parallelised and doing direct path reads bypassing the buffer cache.
In 11gR2, you might also find full table scans of large tables becoming direct path reads.
See this thread : https://forums.oracle.com/thread/2552571
Hemant K Chitale

Tuning top wait events

hi,
why does the following wait event occurs.how to tune these wait events
control file parallel write and direct path load
With Regards
Boo

"control file parallel write" event occurs while the session is writing physical blocks to all control files.
This happens when:
* The session starts a control file transaction (to make sure that the control files
are up to date in case the session crashes before committing the control file
transaction)
* The session commits a transaction to a control file
* Changing a generic entry in the control file, the new value is being written to all
control files
The wait time is the time it takes to finish all writes to all control files
To reduce this wait event, you can decrease the number of your control files (if this number too high) or place your control files to faster disks.
"direct path load" event occurs when a session waits for completion of direct load operations to database files. For example, if you are using SQL*Loader direct path load operation or import dump file made in direct mode.
To reduce this wait event, you can also place your control files to faster disks.

Top Wait Events

hi gurus,
3 node rac 10.2.0.4 serving a packaged application.
Top 5 timed events in awr shown as
Event=CPU time
     Waits=
Time(s)=1,950
Avg Wait(ms)
% Total Call Time=45.3
Wait Class
Event=gc cr multi block request
     Waits= 6,551,055
Time(s)= 1,396
Avg Wait(ms)=0
% Total Call Time=38.9
Wait Class= Cluster
Event=db file scattered read
     Waits= 186,295
Time(s)= 719
Avg Wait(ms)=4
% Total Call Time=18.2
Wait Class= User I/O
Event=db file parallel read
     Waits= 43,383
Time(s)= 241
Avg Wait(ms)=6
% Total Call Time= 5.9
Wait Class= User I/O
Event=      log file sync
     Waits= 71,064
Time(s)= 83
Avg Wait(ms)=1
% Total Call Time= 3.1
Wait Class= Commit
db_block_size=8KB
db_file_multiblock_read_count = default setting of 128
question:
are the high wait values of gc cr multi block request and db file scattered read are due to db_file_multiblock_read_count?
if that's the case, is there a way to find optimum value for db_file_multiblock_read_count?
or any other findings please?
experts, appreciate your valuable help
thanks in advance,
charles

user570138 wrote:
there are queries going for the full table scan with outer joins (milliion of records). those are the same sqls at the top of "sql order by cluster time"in awr with high CPU utilization.
any way to fine tune the instance to reduce the "gc cr multi block request"
apart from changing the code as the code belongs to a package based application please?
Do you have a performance problem ?
You are doing some large tablescans; these are (probably) the root cause of the gc cr multiblock read, the db file scattered reads, and the CPU, but if the queries are necessary and the execution paths are the best that can be done then maybe you just have to recognise that the resource you're using is reasonable for the queries you have to run.
Otherwise
<ul>
(a) can you find a more efficient access path for any of these queries
(b) can you make sure that all these queries run on the same node so that you get some benefit from node-affinity (possibly the object(s) will be remasted to that single node) and reduce the interconnect traffic.
</ul>
Regards
Jonathan Lewis
http://jonathanlewis.wordpress.com
http://www.jlcomp.demon.co.uk
To post code, statspack/AWR report, execution plans or trace files, start and end the section with the tag {noformat}{noformat} (lowercase, curly brackets, no spaces) so that the text appears in fixed format.
"Science is more than a body of knowledge; it is a way of thinking"
Carl Sagan

Top 5 wait event

I need some guidance on my AWR 5 top wait events
I have 10gR2 on Solaris 9.
The top 5 events in my AWR (ran hourly) always contain the following: (not necessary in order)
CPU time
control file parallel write
dbfile parallel write
log file parallel write
log file sync.
Is this an indication of an undersized log buffer ?
My value for log buffer is 14,258,176
I have 4 CPUs
I'd appreciate any help

Hi!
I do have the same problem and trying to figure it out
Top 5 Timed Events Avg %Total
~~~~~~~~~~~~~~~~~~ wait Call
Event Waits Time (s) (ms) Time Wait Class
CPU time 932 71.3
reliable message 2,828 509 180 38.9 Other
control file parallel write 8,759 300 34 23.0 System I/O
db file parallel write 19,813 238 12 18.2 System I/O
control file sequential read 65,435 193 3 14.7 System I/O
share with me, your thoughts
Ravi

How do I the interpret "Disk file operations I/O" wait event?

I have a large and very busy batch database. All of a sudden the "Disk file operations I/O" wait event is in the top 5 in AWR.
The manual page isn't very helpful:
http://download.oracle.com/docs/cd/E11882_01/server.112/e17110/waitevents003.htm#insertedID40
Disk file operations I/O
This event is used to wait for disk file operations (for example, open, close, seek, and resize). It is also used for miscellaneous I/O operations such as block dumps and password file accesses.
So here is my question: What exactly is going on when I see this wait event? Why doesn't it show up as one of the other I/O events? Can I make it go away? Should I make it go away?
DR

sb92075 wrote:
All of a sudden the "Disk file operations I/O" wait event is in the top 5 in AWR.Top wait event
In EVERY Top Wait Event list, one wait event will ALWAYS be on top as #1; by definition of the list.
Simply because any item, even #1, appears on this list does not mean this is a problem & needs to be fixed.
If the Top Wait Event accounts for only 5 seconds out of a 1 hour sample,
then reducing it to ZERO won't measurably improve overall application performance.
The actual Time Waited is required to determine if it is a problem or not.It's taking 20% of time in a 15 minute sample. Anything that takes 20% of deserves to be understood....So: What actually causes it?
DR

What is ges reusing os pid wait event

What is wait event "ges reusing os pid". In our RAC environment it is one of the top wait events. How to minimze it.

This is a wait event in Oracle 10g for Global Enqueue Services (ges) waiting on an operating system process id (os pid).
How to resolve this issue? I checked the bug list on Metalink and there is a patch set for the issue that may help.
Question: what version and patch release are you running for Oracle RAC?
Also, you probably want to tune your public network and private interconnects between the nodes in your Oracle RAC cluster.
Regards,
Ben Prusinski
http://oracle-magician.blogspot.com/

CXPACKET Wait events

Hi,
I am facing wait event CXPACKET in top wait events. we are using SQL Server 2012. the DOP is set to 1.
Please suggest how to resolve this
REgards

Hi,
I am facing wait event CXPACKET in top wait events. we are using SQL Server 2012. the DOP is set to 1.
Any reaosn why DOP is set to 1.Have you tested this scenario.Instaed of first moving for resolving CXPACKET wait which is not a acctual reason it is just a symptom .It happens when various parallel threads are waiting to synchronize after doing the
task.What is other major wait stats you can see.Can you paste output of below query here
--By Jnathan Kehayias
SELECT TOP 10
wait_type ,
max_wait_time_ms wait_time_ms ,
signal_wait_time_ms ,
wait_time_ms - signal_wait_time_ms AS resource_wait_time_ms ,
100.0 * wait_time_ms / SUM(wait_time_ms) OVER ( )
AS percent_total_waits ,
100.0 * signal_wait_time_ms / SUM(signal_wait_time_ms) OVER ( )
AS percent_total_signal_waits ,
100.0 * ( wait_time_ms - signal_wait_time_ms )
/ SUM(wait_time_ms) OVER ( ) AS percent_total_resource_waits
FROM sys.dm_os_wait_stats
WHERE wait_time_ms > 0 -- remove zero wait_time
AND wait_type NOT IN -- filter out additional irrelevant waits
( 'SLEEP_TASK', 'BROKER_TASK_STOP', 'BROKER_TO_FLUSH',
'SQLTRACE_BUFFER_FLUSH','CLR_AUTO_EVENT', 'CLR_MANUAL_EVENT',
'LAZYWRITER_SLEEP', 'SLEEP_SYSTEMTASK', 'SLEEP_BPOOL_FLUSH',
'BROKER_EVENTHANDLER', 'XE_DISPATCHER_WAIT', 'FT_IFTSHC_MUTEX',
'CHECKPOINT_QUEUE', 'FT_IFTS_SCHEDULER_IDLE_WAIT',
'BROKER_TRANSMITTER', 'FT_IFTSHC_MUTEX', 'KSOURCE_WAKEUP',
'LOGMGR_QUEUE', 'ONDEMAND_TASK_QUEUE',
'REQUEST_FOR_DEADLOCK_SEARCH', 'XE_TIMER_EVENT', 'BAD_PAGE_PROCESS',
'DBMIRROR_EVENTS_QUEUE', 'BROKER_RECEIVE_WAITFOR',
'PREEMPTIVE_OS_GETPROCADDRESS', 'PREEMPTIVE_OS_AUTHENTICATIONOPS',
'WAITFOR', 'DISPATCHER_QUEUE_SEMAPHORE', 'XE_DISPATCHER_JOIN',
'RESOURCE_QUEUE' )
ORDER BY wait_time_ms DESC
Please mark this reply as the answer or vote as helpful, as appropriate, to make it useful for other readers

Latch: session allocation in top 5 wait events

Hi,
This wait event session is coming in the top 5 timed events
How to proceed for solving this issue?
This database is only being used for migration activities currently,which means a lot of imports going on.
PFB the Top 5 Timed Foreground Events
++++++++++++++++++
Event     Waits     Time(s)     Avg wait (ms)     % DB time     Wait Class
DB CPU          8,437          57.95
latch: session allocation     2,035,326     3,671     2     25.22     Other
wait list latch free     243,511     2,448     10     16.81     Other
direct path write     504,262     363     1     2.49     User I/O
log file sync     39,396     156     4     1.07     Commit
++++++++++++++++++
Cheers,
Kunwar

user9131570 wrote:
This database is only being used for migration activities currently,which means a lot of imports going on.
PFB the Top 5 Timed Foreground Events
++++++++++++++++++
Event     Waits     Time(s)     Avg wait (ms)     % DB time     Wait Class
DB CPU          8,437          57.95
latch: session allocation     2,035,326     3,671     2     25.22     Other
wait list latch free     243,511     2,448     10     16.81     Other
direct path write     504,262     363     1     2.49     User I/O
log file sync     39,396     156     4     1.07     Commit
++++++++++++++++++
A little more information would help - you could start with the Load Profile section, and the summary of the number of sessions, CPU, and memory. Also OSSTATs might be helpful and the Time Model stats. The version of Oracle might also be relevant.
See the note below about using the "code" tag and preview option.
Regards
Jonathan Lewis
http://jonathanlewis.wordpress.com
http://www.jlcomp.demon.co.uk
To post code, statspack/AWR report, execution plans or trace files, start and end the section with the tag {noformat}{noformat} (lowercase, curly brackets, no spaces) so that the text appears in fixed format.
There is a +"Preview"+ tab at the top of the text entry panel. Use this to check what your message will look like before you post the message. If it looks a complete mess you're unlikely to get a response. (Click on the +"Plain text"+ tab if you want to edit the text to tidy it up.)
+"Science is more than a body of knowledge; it is a way of thinking"+
+Carl Sagan+

Enq: TX - row lock contention in TOP 5 wait event

DB version:11.1.0.7.0
I am having enq: TX - row lock contention in top 5 wait event.
AWR analyze period - 9-10(pm). During this time only one sql loader is running to insert the data. No other job are running. So there is no chance of other session blocking this session. is there any chance of row lock contention happen by same session.
SQL> SELECT INDEX_NAME,INDEX_TYPE,UNIQUENESS FROM DBA_INDEXES WHERE TABLE_NAME='DATA_DATA';
INDEX_NAME INDEX_TYPE UNIQUENES
CIDX BITMAP NONUNIQUE
VIDX BITMAP NONUNIQUE
Thanks.

SQL> SELECT INDEX_NAME,INDEX_TYPE,UNIQUENESS FROM DBA_INDEXES WHERE TABLE_NAME='DATA_DATA';
INDEX_NAME INDEX_TYPE UNIQUENES
CIDX BITMAP NONUNIQUE
VIDX BITMAP NONUNIQUEYou have bitmap indexes here on a table being inserted into. Bitmap Indexes are another source of lock(and deadlock) in OLTP application. You said that the SQLloader was the unique active program but may be you are also triggering another procedure after the load. Procedure in which you might be using also automomous transactions and so on...
Check first if your table is subject to DML operation in a a multi-user concurrent accesss and in which case you have to get rid of those bitmap indexes
http://hourim.wordpress.com/2011/03/14/deadlock-%e2%80%93-part-1-bitmap-index/
Best regards
Mohamed Houri
www.hourim.wordpress.com

Top 5 event class slave wait

Dear Support,
I run the AWR report. It show the top 5 event is class slave wait 70second then follow by CPU time. May i know what is class slave wait event. Thanks
Regard
William

10g is a marketing label not a version number.
SELECT * FROM v$version;Don't describe the AWR report part your are seeing ... Ctrl-C then Ctrl-V ... and post it between tags so we can read it.

Interpret DB CPUwait event (top 5 wait event AWR)

Hi,
Can anyone tell me how to read the table below especially the "DB CPU" section,
Is it right to say that 41.71% of time was consumed waiting for CPU?? this is urgent
Event                     Waits           Time(s)      Avg wait (ms)      % DB time      Wait Class
db file sequential read      300,835      1,483           5           58.42           User I/O
DB CPU                     1,059                     41.71
reliable message           9,499           18           2           0.71           Other
PX Deq: Slave Session Stats      6,506           11           2           0.43           Other
gc cr grant 2-way           26,218           6           0           0.25           Cluster

user589420 wrote:
Hi,
Can anyone tell me how to read the table below especially the "DB CPU" section,
Is it right to say that 41.71% of time was consumed waiting for CPU?? this is urgent
Event                           Waits Time(s) Avg wait (ms) % DB time Wait Class
db file sequential read       300,835    1,483              5      58.42 User I/O
DB CPU                                   1,059                     41.71
reliable message                9,499       18              2       0.71 Other
PX Deq: Slave Session Stats     6,506       11              2       0.43 Other
gc cr grant 2-way              26,218        6              0       0.25 Cluster
When posting information to the forum that includes critical spaces, like the above, use a { code } tag (without spaces) before and after the information.
I do not understand why this question is an urgent problem.
It is incorrect to state that 41.71% of the time was consumed waiting for the CPU. When an Oracle process is running on the CPU, it is officially not waiting. It causes a bit of confusion having the CPU time consumed listed among the top 5 wait events, but as long as you understand why it is in the top 5 list, it almost makes sense for it to be included.
The DB CPU statistic is listed as 1,059 seconds. If the duration of this report is 1 hour, that is 3,600 seconds of total time. If there is a single CPU in the server, there are 3,600 CPU seconds available in the time period, indicating that the server's CPU on average was 29.4% busy. If there were 12 CPUs in the server, there were 43,200 CPU seconds available in the time period, indicating that on average the CPUs were 2.5% busy. Does this mean that there was a problem, or was this OK, or is there not enough information? Just because on average the CPUs are not busy, that does not mean that there were not periods of intense CPU competion, where in fact there was a temporary shortage of available CPU time for processing.
The DB Time statistic is supposed to be an indication of work performed by the instance on behalf of the user sessions. It is the accumulation of CPU time consumed by foreground sessions plus the accumulated sum of all non-idle wait events consumed by foreground sessions. Blog articles that might be of interest to you:
http://hoopercharles.wordpress.com/2010/01/13/working-with-oracle-time-model-data/
http://hoopercharles.wordpress.com/2010/02/05/faulty-quotes-6-cpu-utilization/
Charles Hooper
Co-author of "Expert Oracle Practices: Oracle Database Administration from the Oak Table"
http://hoopercharles.wordpress.com/
IT Manager/Oracle DBA
K&M Machine-Fabricating, Inc.

( name-service call wait   ) event   is amoung the top 5 wait events

10.2.0.3 2node- RAC
When I check the wait event in the active session wait event the sessions can be see that they are blocked by the LMON process. The wait event appears only on one node.
It is very hard to find information on the net or on the metalink.
How could I overcome this wait event ?
select program, type from v$session
     where sid in
     (select blocking_session from v$active_session_history
               where event like 'name-service%'
     and rownum < 1000 )
oracle@dbokyanus1 (LMON) , BACKGROUND
Top 5 Timed Events Avg %Total
~~~~~~~~~~~~~~~~~~ wait Call
Event Waits Time (s) (ms) Time
db file sequential read 4,331,507 15,218 *4* *31.2*
CPU time *8* *17.5*
log file sync 1,462,968 5,513 *4* *11.3*
name-service call wait 72,058 4,545 *63* *9.3*
SQL*Net message from dblink 4,197 4,047 *964* *8.3*

oceanic-815 wrote:
10.2.0.3 2node- RAC
When I check the wait event in the active session wait event the sessions can be see that they are blocked by the LMON process. The wait event appears only on one node.
It is very hard to find information on the net or on the metalink.
If I use Google to find information about this
I ran into this page.
There is indeed not much to find about this wait-event, other than there is probably some sort of network communication problem between the two nodes.
Maybe the best solution is to raise a S/R at Oracle support

Latches in top 5 wait events

Similar Messages

Maybe you are looking for