Concurrent manager using lots of cpu & memory

In the last few days, I have noticed the load on the production server as follows:
load averages: 4.66, 4.86, 4.92 14:34:16
333 processes: 327 sleeping, 6 on cpu
CPU states: % idle, % user, % kernel, % iowait, % swap
Memory: 32G real, 7591M free, 23G swap in use, 14G swap free
PID USERNAME LWP PRI NICE SIZE RES STATE TIME CPU COMMAND
20045 appsdapo 1 0 0 4073M 4029M cpu/2 165.0H 12.49% f60webmx
1854 appsdapo 1 10 0 4082M 4060M cpu/0 240.0H 12.48% FNDLIBR
2504 appsdapo 1 10 0 1872M 1869M cpu/19 1980.2 12.46% FNDLIBR
2496 appsdapo 1 0 0 1876M 1873M cpu/1 1957.5 11.68% FNDLIBR
18641 appsdapo 1 10 0 1512M 1509M cpu/16 1080.3 8.64% FNDLIBR
I have come across Metalink note ID 114380.1 which says identify the C.M but how do I relate the FNDLIBR processes above with the specific C.Managers.
Thanks

Hello,
Thanks for the Post. This post helped me to solve my production issue.
My production application started hanging suddenly and no clue why it is behaving like this.
The symptom was FNDLIBR processes were on top cpu consumer on application server.
The database server also experienced heavy system i/o and cpu usage.
The database base log switches every minutes (500 MB) and there are no concurrent requests running.
No concurrent request errors. No OS errors (Cleaned the GNOME and restarted). No database errors.
But in sql area i saw lot of fnd service requests keep coming.
At the end the application server hangs because of high cpu and it caused the production down.
The problem was one of Our System Administrator disabled one responsibility with scheduled request.
One interesting thing we noticed, even you put the request in question Hold, it will start running while you restart the concurrent request.
You have to cancel the concurrent request and restart the concurrent manager can solve the issue.
Thx
Pouler

Similar Messages

Running report in concurrent manager using unix shell script to create PDF

Hi,
I need help urgently, we are in the process of migrating from 10.7 to 11i. My problem is we have a report created in 10.7 that is ran through concurrent manager using shell scripts(host) and the output is stored as .pdf, in 10.7, it works perfectly. Now I am also doing it in 11i but it gives some error. The shell script from 10.7 is: r25runm module=$XXX_TOP/srw/test.rdf \
userid=$1 batch=yes \
desformat=pdf destype=file \
desname=$XXX_TOP/outbound/test.pdf
I change the shell script in 11i as follows:
ar60runb report=$XXX_TOP/reports/US/test.rdf \
userid=$1 batch=yes \
desformat=pdf destype=file \
desname=$XXX_TOP/outbound/test.pdf
I checked your metalink, and I am confused which is the right executable I should use, is it ar60runb or ar60run or rwrun60b or rwrun60c or rwrun60b, which is which? how many parameters do I have to include?, some documents are saying I have to use the 4 parameters-orauser/pwd, userid,username, request_id, others one.Which parameter/s go/es with what executable. We are using HP-UX server 64 bit. In 11i, when I run in concurrent manager it gives me an error:The executable file /chdev/fd11/u00/fd11appl/xxx/1.0/bin/test for this concurrent program cannot be executed.
Contact your system administrator or support representative. Verify that the execution path to the executable file is co.
I have checked Metalink and follow the directions, created a link fndcpesr, check and set the permissions, etc. Play around with the executables ar60runb or rwrun60 etc., the parameters. The ar60* or rwrun60* executable permissions are -rwxr-xr-x and $XXX_TOP/fnd/../fndcpesr is -rwxr-xr-x. Also, why is ar60* executables located in $FND_TOP/bin, whereas rwrun60* executables are in $ORACLE_HOME/bin? Please help, I need an answer urgently ,I have to complete this task before Tuesday 9/16/2003 for our migration deadline. Thank you very much.

I have already fixed the problem, TY anyway

Dock uses lots of CPU time

I find Dock uses ~140% CPU time sometimes, I can't see a cause for it to do so. It usually stops of its own accord after a random amount of time. Killing it doesn't help, launchd restarts it and it goes back to using lots of CPU.
I have parallels install and I have the version which has the fix (6.0.12106) and I have also tried the work around (no luck obviously).

Never mind, I figured it out by myself. On this particular machine I had a 3rd party plug-in installed (SRS iWow 1.1.8). Removing that brought the CPU usage down to 0.1%. Usually I don't install 3rd party add-ons. This would seem to be a good policy.
Steve

Gnome pulseaudio using lots of shared memory

Out of curisoty I did the following ls -l /dev/shm and I saw the following files:-
-r-------- 1 peter users 67108904 Mar 23 10:40 pulse-shm-3021221787
-r-------- 1 peter users 67108904 Mar 23 10:24 pulse-shm-3538772094
-r-------- 1 peter users 67108904 Mar 23 12:27 pulse-shm-3584709017
-r-------- 1 peter users 67108904 Mar 23 10:24 pulse-shm-419282190
Just wondering why pulseaudio needs such large files? Is it possible to reduce this?
Many thanks in advance

Hello,
Thanks for the Post. This post helped me to solve my production issue.
My production application started hanging suddenly and no clue why it is behaving like this.
The symptom was FNDLIBR processes were on top cpu consumer on application server.
The database server also experienced heavy system i/o and cpu usage.
The database base log switches every minutes (500 MB) and there are no concurrent requests running.
No concurrent request errors. No OS errors (Cleaned the GNOME and restarted). No database errors.
But in sql area i saw lot of fnd service requests keep coming.
At the end the application server hangs because of high cpu and it caused the production down.
The problem was one of Our System Administrator disabled one responsibility with scheduled request.
One interesting thing we noticed, even you put the request in question Hold, it will start running while you restart the concurrent request.
You have to cancel the concurrent request and restart the concurrent manager can solve the issue.
Thx
Pouler

Spotlight endlessly indexing and mdworker process using lot of cpu

Hi, I just did Archive&Install from Tiger to Leopard, applied 10.5.1 and now...
...since THREE DAYS Spotlight is indexing giving strange values for time remaining (2.134.545.673 hours).
Like I found in other threads, I have the mdworked process absorbing a lot of cpu power and, from Console, maybe it's crashing repeatedly. I made private of lots of folders but the problem is still there.
Any other idea?
How can I find which file/application is freezing mdworker and making Spotlight so unuseful?
Thanks in advance
Tampano

I don't know. How much data is on your HD? Simplest thing to do is use the Terminal command mdutil to erase the current indexes and rebuild them. This should do it:
*sudo mdutil -E /Volumes/(asterisk)*
Launch the Terminal.app in /Applications/Utilities/, copy & paste that command into the window, hit the return key, enter your admin password at the Password: prompt (carefully because nothing shows up on the screen), and hit the return key again. That should erase the indexes and rebuild them.

Multiple instances of Internet Explorer in task manager using 100,000 KB+ memory

Hi! I recently got a virus on my computer that causes multiple instances of explorer.exe and iexplorer.exe to appear in my processes tab of task manager. Over the course of about a minute, the memory usage of each begins to climb rapidly until all of my
memory is being used by these. These do not appear in my applications, nor do they actually open up an internet explorer window. When I end task on them, another one appears about a minute later to take its place and do the same thing again. Additionally,
when I click shut down to turn my computer off, various advertisements flash on my screen for a few seconds before the computer begins to shut down. When the computer is started up, a window appears for a split second that contains only blackness. I have never
seen this window appear before the virus was obtained. When I end task on all my explorer.exe and iexplorer.exe applications at the same time, my desktop vanishes, but it also stops those applications from reappearing. This virus has caused my computer
to run unbearably slow. I have tried Spybot S&D, Exterminateit, AVG, Roadkill, and other methods and nothing seems to detect this or cure it. I am not that good with computers, but I am dying for help on the matter! If you can assist, please let me know
what I can do with detailed instructions for my sake! Thank you!!!!!!

See and follow Andy's excellent advices.
In addition...
See: "How to
Reset Internet Explorer settings".
If the issue persists...
1. Please download and run the
Microsoft Safety Scanner tool (freeware).
2. Reboot your PC.
3. Please download and update Malwarebytes Anti-Malware Free
4. Reboot Windows in
Safe Mode and run a Full Scan with Malwarebytes Anti-Malware.
5. Download and Run AdwCleaner (Freeware)
6. Reboot and Test.
Best of Luck,
Vincenzo Di Russo
Microsoft® MVP Windows Internet Explorer, Windows & Security Expert - since 2003.
Moderator in the Microsoft Community and TechNet Forums
My MVP Profile

Firefox 3.6.12 is using lots of CPU when shutting down, and doesn't ever die

It may have something to do with changing settings, but on the closing phase of restarting Firefox 3.6.12, one of my two CPUs get's used, CPU running at about 50% to 60%. It doesn't seem to stop until I kill it from task manager. I haven't been able to wait more than 10 minutes for it.

@cor-el
Thanks for clearing up my misunderstanding. As i understood now it is in fact only a "DOM storage data" topic.
Quote from betterprivacy:
''"Disabling DOM storage (instead of deleting the file on shutdown) causes a website script to throw an exception. This means that the execution of the script is aborted immediately, not just the DOM storage instruction is ignored. None of the following script instructions would be executed, even if those instructions have nothing to do with DOM storage. That is the reason why -with DOM storage disabled- sites like CNN.com will cease to play videos."''

IPhoto uses lots of CPU time while idle

I run iPhoto '08 (7.0.2) on a G4 iMac that I use as a photo and music server. Today I was using the computer to rip some CDs and scan the covers and the computer seemed sluggish. I checked Activity Monitor and saw that iPhoto was using between 20 and 60% of the CPU time even though it was hidden and hadn't been active for at least a day. When I activated iPhoto it went back down to 1% or less. I hid it and it remained at 1% or less for a while, but after about 15 minutes it shot back up to 20-60%. Any ideas as to what might be going on?

I seem to find that both iPhoto and Safari grow in virtual memory used and CPU used as time goes by so I try to restart each of them every couple of days and he whole computer at least once a week (no it shouldn't be necessary -- but at least it is still better than windows)
Larry Nebel

ITunes uses lots of CPU time when idle

When iTunes is just sitting there with my Music Library displayed, Activity Monitor shows it's using a pretty consistent 22-25% of CPU time. I look at Activity Monitor once in a while and I've never noticed this before. I don't know if it started when I upgraded to iTunes 7.4.2 or not. Other Macs I have access to use close to 0%. I have repaired permissions, checked the hard drive, and waved the other dead chickens over it without any change. Obviously this soaks up a lot of machine resources. Any clues or ideas would be welcome.
Steve

Never mind, I figured it out by myself. On this particular machine I had a 3rd party plug-in installed (SRS iWow 1.1.8). Removing that brought the CPU usage down to 0.1%. Usually I don't install 3rd party add-ons. This would seem to be a good policy.
Steve

Plotting on more than one CWGraph components uses lots of cpu resource.

The cpu usage is over 95%. When we send some triangle waveform data over a DS socket, then receive the data from another appliction and try to plot the data on a CWGrahp component, the graph shows we loss some data points.points in the

Hi Juan:
We are trying to develop a data acquisition system. We need to pass 16 values every 8.33 ms by using DS Socket and display the 16 values in 16 CWGraph components.
I wrote two applications by modifying the DS Writer/Reader With Attributes examples.
To change the interval of the Sender, you have to change the interval of Timer1 in its properties in VB IDE.
To change the number of charts ploted on the CWGraph components, you have to comment or un-comment out the corresponding line in the Receiver VB code.
I run them on a Pentium 3, 600MHz, PC, the test result is:
1. The Receiver can not receive all the 16 values when the sender sends them in less than every 10 ms.
There are some packets lost during the transmission.
2. Whi
le the Receiver not plotting any chart, it can receive all the 16 values corrctly in all the packets when the Sender sends a packet in every 11 ms.
3. If the Receiver plots on one chart, it can receive all the 16 values corrctly in all the packets when the Sender sends a packet in every 12 ms.
4. If the Receiver plots on 16 charts and the Sender sends a packet in less than every 84 ms, the CPU usage will be 100% and the controls, such as lables, on the receiver GUI will not be updated corrctly.
5. If the Sender sends a packet in greater every 85 ms, the Receiver can receive the data and plots them on 16 charts corrctly.
The test results do not meet our requirement and my application is aimed to be used on some less powerful laptops. I hope it is not because the performance limit of the DS Socket and CWGraph.
I was wondering whether we can do something to increase the performance.
Pls find the attached source code files.
Thank you!
Attachments:
VerSentToNI.zip ‏36 KB

Concurrent Manager Issue in Oracle Apps 11i on MS Windows 2003

Hi,
We have Oracle Apps 11i (11.5.10) on Windows 2003 running Oracle Financilas and Supply Chain.
From this morning onwards, concurrent requests are not getting completed. All requests are showing Running Normally. I checked the server (Single Node 4 GB Memory) and CPU usage is more than 90%. Task Manager is showing 50% CPU usage for SYSTEM (Image Name). Once the CPU usage for SYSTEM(image name) goes beyond 40% then the server is becoming slow and the problem starts.
When I stop the Concurrent Manager it does not stop and I have to End Process the FNDLIBR process manually.
I have recreated the concurrent manager using CCMSETUP and also ran cmclean.sql. Nothing is helping.
System is Live since 2006 and the issue started only today.
Please suggest any solutions as this is happening in our Production system.
Thanks.
Thiru

Thiru,
Which concurrent manager consumes the CPU usage? Are you able to identify it?
Any errors in the CM log file? Did you try to start the CM with diag=y and see if any errors are reported in the log files?
What happen to the CPU usage when you stop the CM?
I assume that you are running Purge Concurrent Manager/Requests on regular basis, so this should not be an issue here.
Please review the following notes, and see if it helps.
Note: 114380.1 - Concurrent Manager Processes Taking Lots of CPU
https://metalink2.oracle.com/metalink/plsql/ml2_documents.showDocument?p_database_id=NOT&p_id=114380.1
Note: 264752.1 - FNDLIBR Consuming Memory Causing CPU Utilization To Increase Daily
https://metalink2.oracle.com/metalink/plsql/ml2_documents.showDocument?p_database_id=NOT&p_id=264752.1
System is Live since 2006 and the issue started only today.I assume no changes have been done recently, correct?
Regards,
Hussein

CPU/memory usage of AIR application under Mac 10.4.10

As my AIR application uses a lot of CPU/memory under Mac
(seems better in Windows), I decide to do some basic testing. I
create several applications with different level of complexities.
Below is a table of CPU/memory usage when the application is simply
sitting there. A brief description of the application can be found
in the end of each line.
Here are some results:
[1] Even for an empty application, CPU usage is not zero
(most non-AIR application has 0 CPU usage when doing nothing). Also
the memory usage is huge, comparable to Finder on my computer.
[2] Empty application has 7 threads running.
[3] Indetermined progress bar costs a lot of CPU; and I guess
any other animation costs a lot of CPU too. But if they are
invisible they are fine.
[4] running timer costs CPU.
PID COMMAND %CPU TIME #TH #PRTS #MREGS RPRVT RSHRD RSIZE
VSIZE application
10093 adl 1.8% 0:01.85 7 100 292 15.0M 13.8M+ 22.8M 444M
empty
10094 adl 1.9% 0:01.55 7 100 299 15.2M 14.0M+ 23.1M 444M
label
10097 adl 1.9% 0:01.48 7 100 302 15.3M 13.4M 23.7M 445M
progress bar
10095 adl 5.7% 0:02.40 7 100 302 15.4M- 13.2M 23.5M 445M
progress bar (indetermined)
10125 adl 15.9% 0:03.66 7 100 304 15.7M- 13.4M 23.9M 445M 5
progress bar (indetermined)
10126 adl 29.7% 0:10.39 7 100 325 16.7M+ 13.4M+ 24.9M+ 445M
10 progress bar (indetermined)
10128 adl 1.9% 0:01.34 7 100 324 16.1M 13.7M 24.3M 445M 10
progress bar (indetermined), visible=false
10136 adl 1.8% 0:00.76 7 100 323 16.5M+ 13.9M+ 24.7M- 450M
Empty datagrid
10147 adl 1.9% 0:01.60 7 101 328 16.1M 13.8M+ 24.5M- 447M
TabNavigator with two empty VBox children
10138 adl 1.8% 0:01.59 7 100 372 19.4M 13.9M 27.7M 453M
datagrid with 10000 simple rows
10131 adl 3.1% 0:02.38 7 110 348 25.7M 19.3M 34.8M 458M 10
Empty native windows
10143 adl 1.9% 0:00.99 7 100 296 15.1M 13.9M+ 23.2M- 444M
Timer, delay = 100ms, calling an empty function
10145 adl 4.1% 0:02.35 7 100 298 15.0M 14.0M+ 23.2M- 444M
Timer, delay = 100ms, calling function update label position
As to my own application, it's much bigger than these simple
ones. The CPU usually stablizes to 20% and VSIZE is about 1G. There
is no timer running. When I leave the application open for a few
hours, the application will eventually freeze and I have to kill
it.
I am wondering if there are any articles or documents
explaining AIR's CPU/memory usage. And also if AIR is planning to
improve the performance in the next release.

James, thank you for the very fast answer! I did various tests on the weekend on my computer and computer of my friend.
And I can say, that Debug flag doesn't affect memory consumption. They still eat memory. My steps to reproduce are as follows:
(Windows 7 32 bit)
Photoshop CS5:
1) set HKEY_CURRENT_USER\Software\Adobe\CSXS2Preferences\PlayerDebugMode to 0
2*) check HKEY_LOCAL_MACHINE\SOFTWARE\Adobe\CSXS2Preferences (just in case)
3) Compile HelloPhotoshop example to ZXP and install via Extension manager
4) restart computer
5) Run Photoshop CS5 - Memory consumption: 100 096 KB
6) Open HelloPhotoshop and close it 10 times
7) Memory consumption: 155 840 - 5,574 MB per one close/open
Maybe a garbage collector is expected to run and free this memory, but it never happened in my case. The memory pool was increasing to 1 GB when I was working with one of my panels - with no documents open in Photoshop
In InDesign I tried the same thing with HelloInDesign ( I had to close the panel completely and then open with Window | Extensions | HelloInDesign)
Memory consumption on start: 118 516
Memory consumption after 10 open/close cycles: 197 516
It's 7,9 MB per one close/open
It's the simpliest panel test. When I'm using AIR libraries the memory usage is increasing a lot faster.

Helpd takes a lot of CPU

What is this process ?
It's spawned quite often, and uses a lot of CPU & memory.

Helpd updates all of the Application help files on your Mac.
However, I hate having helpd launching repeatedly and seemingly randomly, taking up 1 gig of RAM, and using so much CPU time that the fans go loudly full blast on my MacBook Pro. I prefer having my MacBook Pro be quiet.
Upon investigation, I found that helpd is controlled by a System LaunchAgent plist which is run on startup:
/System/LaunchAgents/com.apple.helpd.plist
When examining the contents of this plist using XCode, what you find is that:
1. helpd is run when it is loaded by the system at startup or login
2. helpd is run on every Sunday at 2:30 AM
3. the launch agent watches the following folders and runs helpd whenever the folders change:
    /Applications/
    /Applications/Utilities/
The biggest problem is the watch folders. They change whenever you open them (i.e. their modified date changes when opened). They change when you install a new application. Many applications change their Application folder content when run. etc. etc.
(Lingon is a simplified launch agent and launch daemon editor. Lingon 3, by the way, cannot manage anything other than your own user's launch agents and launch daemons. It is limited by Apple's new App Store rules. Lingon 2 can open up system launch agetns and launch daemons. But it is missing many of the plist fields revealed by using XCode. For example, it will show the /Applications/ watch folder. But there are TWO watch folders - the second one is /Applications/Utilities/ Thus, Lingon may cause problems by modifying the system launch agents and daemon plists by missing several fields. I don't recommend Lingon other than to create your own user launch agents and daemons - with Lingon 2 being a much more capable version)
The easiest solution is to use Xcode or another plist editor to modify the com.apple.helpd.plist to change the watch folders.
I changed /Applications/ to /Applications/Utilities/ and left the second watch folder /Applications/Utilities/
Since I seldom open /Applications/Utilities/ this causes helpd to launch far less often.
You can also just simpliy eliminate the watch folders. This keeps helpd running on Sundays at 2:30 AM or on startup or login.
When users get rid of many older applications, they see helpd launch less often since these older applications may be the ones modifying the /Applications/ folder.
If you have only newer applications or few applications, you don't run into the helpd problem as often or at all. This is why few people complain about helpd. Only the users who have numerous applications (particularly older ones or ones not bought from the App Store) run into this problem.
Here are the contents of the com.apple.helpd.plist:
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
          <key>EnableTransactions</key>
          <true/>
          <key>EnvironmentVariables</key>
          <dict>
                    <key>NSRunningFromLaunchd</key>
                    <string>1</string>
          </dict>
          <key>Label</key>
          <string>com.apple.helpd</string>
          <key>LimitLoadToSessionType</key>
          <array>
                    <string>Aqua</string>
                    <string>LoginWindow</string>
          </array>
          <key>MachServices</key>
          <dict>
                    <key>com.apple.helpd</key>
                    <true/>
          </dict>
          <key>OnDemand</key>
          <true/>
          <key>Program</key>
          <string>/System/Library/PrivateFrameworks/HelpData.framework/Versions/A/Resourc es/helpd</string>
          <key>RunAtLoad</key>
          <true/>
          <key>StartCalendarInterval</key>
          <dict>
                    <key>Hour</key>
                    <integer>2</integer>
                    <key>Minute</key>
                    <integer>30</integer>
                    <key>Weekday</key>
                    <integer>7</integer>
          </dict>
          <key>WatchPaths</key>
          <array>
                    <string>/Applications/</string>
                    <string>/Applications/Utilities/</string>
          </array>
</dict>
</plist>

Unable to start Concurrent Manager.

Hi group,
Me once again :=D, seems that Ebusiness Suite doesn't want to work properly on Windows environment :=D
I am trying to start Concurrent Manager using "adcmctl.cmd" script, but it is unable to start it, please see below message for this error.
F:\ORAEBS_APPS\inst\apps\EBSTEST_srvoracle05\admin\scripts>adcmctl.cmd start
ECHO is off.
ECHO is off.
Tue 06/01/2010 07:15 PM
ECHO is off.
ECHO is off.
Tue 06/01/2010 07:15 PM
ECHO is off.
ECHO is off.
Setting environment for using Microsoft Visual Studio 2008 x86 tools.
APPSORA.cmd exiting with status 0
Enter the APPS Username/Password: apps/apps
Starting concurrent manager for EBSTEST ...
The OracleConcMgrEBSTEST_srvoracle05 service is starting.
The OracleConcMgrEBSTEST_srvoracle05 service could not be started.
A system error has occurred.
System error 1067 has occurred.
The process terminated unexpectedly.
adcmctl.cmd exiting with status 2
There aren't messages in Concurrent Manager logfile.
I just recreated the service once again using adsvcm.cmd script.
Any help or advice will be really appreciated.
Thanks in advance.
Kind regards,
Francisco Mtz. Oviedo

Hi,
How often do I need to apply this procedure?, just kidding !!!! :=DUntil you migrate to Linux/Unix -- Not kidding :)
It is Windows and I need to bounce everything when I need to solve this kind of issue !!!Yes, unfortunately this is how it goes. I do not like to maintain Oracle Apps on Windows especially when it comes to the CM!
I was getting next error message in Event Viewer for Application:
Error occured: StartICM:Logfile:F:\ORAEBS_APPS\inst\apps\EBSTEST_srvoracle05\logs\appl\conc\log\CM_EBSTEST_srvoracle05.log:The process cannot access the file because it is being used by another process. (0x20)Download this software and see if you can find what other processes are accessing this file and kill it (if possible!).
Process Explorer
http://technet.microsoft.com/en-us/sysinternals/bb896653.aspx
Thanks,
Hussein

Database consuming lot of Physical memory

Hi ,
My database is on version 11.1.0.7.0 and on SUN SOLARIS SPARC.
My server admin just informed me that my database is using lot of physical memory , which i understand is RAM.
I am looking on google also but i am not able to find a way in which i can check on it and see how it can be controlled.
Any help/suggestion would be highly appreciated.
Regards
Kk

There are 2 basic methods that Oracle uses memory.
Statically. Oracle allocates memory (for the SGA) when it starts. This memory remains fixed in size.
Dynamically. In order to service a client, memory is needed for that client session. Oracle dynamically allocates memory for such sessions (called the PGA).
When Oracle memory consumptions grows, it must be dynamically allocated memory. Static memory is just that - static. It does not grow in size.
The usual reason for PGA memory consumption to grow is incorrectly designed and coded bulk processing. A single Oracle server process can easily consume all available free memory on the server as Oracle dynamically increases the size of the PGA of the process running the flawed PL/SQL code.
However, one should not be looking at o/s command line commands to determine Oracle processes's memory utilisation. The output of such commands are often incorrectly interpreted.. as shared memory can be (and often is) included to provide a process's memory utilisation. There are notes on Metalink (mysupport.oracle.com) on the topic and how to correctly use CLI commands to view Oracle process memory utilisation.
An easier, and more accurate, view of Oracle memory utilisation can be obtained from Oracle's virtual performance views.
So, a sysadmin e-mailing a ps (Unix/Linux process listing) showing a particular Oracle process "+using too much memory+" is not really solid enough evidence that memory is being abused. One needs to look closer at the type of memory used by the process.

Concurrent manager using lots of cpu & memory

Similar Messages

Maybe you are looking for