Re: Q: what cause pmon/smon timer wait ?

Hello Ted,

This database has several problems, but they are not the ones you think. The 'pmon timer' and 'smon timer' events are idle waits. These background processes wait on these event when they have no work to do.

Your first problem is that you don't have sufficient freelists on the table to which you are inserting, or possibly have a primary key index on a sequence generated primary that is not a reverse key index. This is probably what is causing your 'buffer busy waits' and 'buffer deadlock' waits.

You also have 'enqueue' waits that are probably a side effect of the lack of process freelists. Also, if your 'latch free' waits are largely on the 'cache buffers chains' latches then that is probably another side effect.

The ratio of you 'log file sync' to 'log file parallel write' waits indicates that your log_buffer is much too big. Drop it back to 160K.

Your 'write complete waits' and 'rdbms ipc reply' waits indicate the either DBWn's write batch is too big, or you are checkpointing too intensively. Make sure that db_files is not much higher than the number of datafiles that you actually have. If you are using incremental checkpoints (db_block_max_dirty_target) don't be too ambitious.

Hope this helps,
Regards,
Steve Adams

http://www.ixora.com.au/

http://www.oreilly.com/catalog/orinternals/

http://www.christianity.com.au/

On Tue, 26 Oct 1999 21:30:25 GMT, tedchyn_at_yahoo.com wrote:

>
>
>Sir, following is a list of v$system_event. what's cause all these
>waits.increasing the number of block buffer, sharepool and sort area
>only has marginal impact on these waits.there are 50 rbs and I resize
>them with large segment and only have marginal impact. I am using oracle
>8.0.5.1 hp with 6 processors(class k) using raw partition. This database
>is use to bench mark a heavily inserting database.
>Could these waits caused by cpu is so much faster than i/o.
>Any insight is appreciated.
>
>ted chyn
>
>
>EVENT TOTAL_WAITS TOTAL_TIMEOUTS TIME_WAITED
>AVERAGE_WAIT
>--------------------------- ----------- -------------- -----------
>------------
>Null event 1 1 401
>401
>latch free 623534 451472 914831
>1.467171
>pmon timer #### 4041 4035 1215290
>300.739916
>process startup 7 0 25
>3.57142857
>rdbms ipc reply 8000 0 82468
>10.3085
>rdbms ipc message 105790 14558 5503890
>52.0265621
>enqueue 157396 2 918233
>5.83390302
>control file sequential read 142 0 0
>0
>control file parallel write 4056 0 8055
>1.98594675
>refresh controlfile command 1 0 0
>0
>free buffer waits 1601 0 622
>.388507183
>buffer deadlock 48272 46659 93475
>1.93642277
>write complete waits 1193 0 14799
>12.4048617
>buffer busy waits 636528 1557 549700
>.863591232
>log file sequential read 14 0 16
>1.14285714
>log file single write 22 0 8
>.363636364
>log file parallel write 558785 0 187438
>.335438496
>log file switch completion 66 0 1126
>17.0606061
>log file sync 2771634 210 2842901
>1.02571299
>db file sequential read 6823 0 1463
>.214421809
>db file scattered read 341 0 242
>.709677419
>db file parallel write 17350 4 108726
>6.26662824
>direct path read 77 0 0
>0
>direct path write 55 0 0
>0
>sort segment request 1 1 101
>101
>instance state change 2 0 0
>0
>smon timer #### 43 39 1193952
>27766.3256
>library cache pin 205 0 30
>.146341463
>file identify 37 0 2
>.054054054
>file open 1261 0 20
>.015860428
>SQL*Net message to client 2818487 0 4344
>.001541252
>SQL*Net more data to client 2 0 0
>0
>SQL*Net message from client 2818486 0 4438999
>1.57495868
>SQL*Net break/reset to client 6 0 0
>0
>
>
>Sent via Deja.com http://www.deja.com/
>Before you buy.
Received on Thu Oct 28 1999 - 07:10:18 CDT