High CPU useage when more than 30 sessions logout/login

Hi guys,
I got this problem for a long time.
Environment : RAC 10.2.0.5 . Oracle enterprise linux 4.8. Java B2C application with dbcp and oc4j dirver.
Server type:Dell R815. 48 cpus 128 G MEM for each node.
Here's many of web application's connect to database and when one of them is down it will restart soon and try to connect to database again.Each one will try to get 50 connections concurrently in the first time.
So , when it happens ,the cpu useage of sys% will grow to more than 70 ,80 ,90% every time.
BTW, it also happens in more than 30 logout concurrently.
Could any one tell me what's the problem is ?
Thank you very much.
Edited by: Longfei Wei on Feb 17, 2011 11:48 PM

Longfei Wei wrote:
Hi guys,
I got this problem for a long time.
Environment : RAC 10.2.0.5 . Oracle enterprise linux 4.8. Java B2C application with dbcp and oc4j dirver.
Server type:Dell R815. 48 cpus 128 G MEM for each node.
Here's many of web application's connect to database and when one of them is down it will restart soon and try to connect to database again.Each one will try to get 50 connections concurrently in the first time.
So , when it happens ,the cpu useage of sys% will grow to more than 70 ,80 ,90% every time.
BTW, it also happens in more than 30 logout concurrently.
Could any one tell me what's the problem is ?
The task of creating a connection and session has to be serialised, and if you start a number of logon or logoff attempts simultaneously you can end up with a storm of activity on the relevant latches that has serious knockon effects. I am a little surprised that you're seeing this with only 30 - 50 connections, but possibly you only have a couple of CPUs.
I made some comments about this on another thread some time ago: Re: Overhead of having high value of processes (or sessions) init parameter
It's possible that you can configure the application server to limit the rate at which it tries to create or terminate connection - e.g. there may be configuration settings to say things like: "create no more than 5 at a time", "wait one second between disconnections". (Oracle has introduced some parameters relating to logoff storms - but perhaps they don't work in your version).
Regards
Jonathan Lewis
http://jonathanlewis.wordpress.com
http://www.jlcomp.demon.co.uk
A general reminder about "Forum Etiquette / Reward Points": http://forums.oracle.com/forums/ann.jspa?annID=718
If you never mark your questions as answered people will eventually decide that it's not worth trying to answer you because they will never know whether or not their answer has been of any use, or whether you even bothered to read it.
It is also important to mark answers that you thought helpful - again it lets other people know that you appreciate their help, but it also acts as a pointer for other people when they are researching the same question, moreover it means that when you mark a bad or wrong answer as helpful someone may be prompted to tell you (and the rest of the forum) what's so bad or wrong about the answer you found helpful.

Similar Messages

Google Earth Plugin and LabVIEW: High CPU usage when adding placemarks

Hi,
I posted this question on stackoverflow earlier this week but feel it might be better suited to the LabVIEW community specifically so I'm reposting here:
I'm writing an application which uses the Google Earth Plugin to display events on the globe. Each event is a single point kml placemark with an icon which is a 3kb png file. Placemarks are uploaded to the plugin as they are received by the software. I am experiencing increasing CPU usage with the number of placemarks that are added.
I have tested displaying a new placemark every second and running until the software running the plugin completely froze (graph attached). The GEPlugin (green trace) stopped responding (i.e. the globe did not respond to the mouse) at around 1200 placemarks added and CPU usage was at ~30%. When the software itself (red trace) froze the plugin was using around 50% CPU and ~3700 placemarks had been added). After the freeze, no new placemarks were added which caused the software to respond (but not the plugin) so I could clear all the placemarks. After the placemarks were cleared from the globe, the CPU usage of the plugin returned to around 5% CPU.
So what I've seen is that GEPlugin CPU usage increases linearly with each kml placemark added. Is this the expected behaviour/ a normal limitation of the plugin? If not is there a more efficient way of adding many placemarks to the globe?
I am using GEPlugin version 7.1.1.1580 (API version 1.010) and LabVIEW 12.0f3
Please see the test results attached. Any input greatly appreciated!
Original stackoverflow post:
http://stackoverflow.com/questions/20994323/google-earth-plugin-with-labview-high-cpu-usage-when-add...
Attachments:
Performance Log 020114_095115.png ‏82 KB

Hello,
I have had a look at your graphs and understood what you are trying to do. To me it seems that as the image gets more complex it gets harde to render which wold likely cause increase in CPU usage resulting in the freeze. I would suggest you try running the program on anoher computer to check on the RAM front of things. If this is a limitation of the GE Plugin then unfortunately I can not do much to help, but if you think this is a problem coming from your LabVIEW code then you can post your code here and I can take a look.

Since I upgraded Firefox, I cannot open more than 1 session at a time. What can I do to correct the problem?

Since I updated Firefox, I can no longer open more than 1 session at a time. How can I correct the problem?
== This happened ==
Every time Firefox opened
== Since last upgrade

Firefox only runs one session at a time, are you talking about not being able to open a 2nd window?
Do you have that problem when running in the Firefox SafeMode?
[http://support.mozilla.com/en-US/kb/Safe+Mode]
''Don't select anything right now, just use "Continue in SafeMode."''
If not, see this:
[http://support.mozilla.com/en-US/kb/troubleshooting+extensions+and+themes]

No more than 1 sessions at a time are allowed. Please wait until open sessions expire.

I have been unable to log into my Actiontec MI424WR router this week, The router returns the message "No more than 1 sessions at a time are allowed. Please wait until open sessions expire." every time I try to log in. This has been going on for days. Even pulling the power to reboot it did no change. I'm the only user in the household that would log into it, so I don't understand how this could occur. How can I get in?
GLN2

I really don't want to do a hard reset as I have a handful of custon port forwarding rules that I had created. I'm not a wiz at that, so it would take me too much time to re-create! I may have no choice. Next time I can get in I will try the backup of the config file. I hope it contains the port forwarding rules.
The backup function should work fine, I have used it to save my Steam port forward rules to my hard drive, after a pin hole reset I was able to restore successfully. I did configure mine to 1 session & ALWAYS logout when I am done. I have never had any problems logging into the router. But in your case you may need more than 1 session allowed.
Are you using a native or 3rd party app to remember your router ID/password? IF you are do not do this, you are "defeating" the built in security of the router
If a forum member gives an answer you like, please give them the Kudos they deserve. If a member gives you the answer to your question, mark the answer as Accepted Solution so others can see the solution to the problem. Thanks !!!
http://forums.verizon.com/t5/Verizon-net-Email/Fix-for-Missing-Inbox-sent-folders-etc-with-Internet-Explorer-11/m-p/647399

High CPU usage when graphics card is put at work

I have a 13inch Air with 2GB RAM and 256GB with Windows 7 Ultimate 64bit. All possible updates have been installed.
I noticed some tearing when trying to move program windows across the screen so I started checking the CPU usage. I saw that it is spiking to 25-35% when moving a window across the screen.When you drag down a window you can see the different copies of the top bar across the screen. Not smooth at all like on a standard PC windows system. Same issue when scrolling through webpages which spikes the cpu
I also noticed a high cpu usage when playing mkv files 25-30% while on an Atom HTPC i get a lower load playing the same file. The Air should be faster than an atom with a nvidia 9400m when playing video files.
I already tried updating the nvidia files to the latest version on their site by hacking the driver inf but this did not resolve the issue.
Has anyone experienced the same or know what could be causing the beavior?
Many thanks

Maybe my initial post was not clear so let me add some clarifications.
I have the problem with Aero disabled
I have the problem on the Apple site which does not have any flash content.
I have already hacked the lastest nvidia drivers to get them on my system which did not change anything. Can you provide a link to some other 'supported' beta drivers?
I am using the new 13 inch air so it has the newer graphics card so I would expect it to be faster than an old HTPC with a 9400m in it. I wasn't stating that my Air had a 9400m.
It would be great if someone you also do the test to check the cpu load:
- when playing a 720 mkv
- when dragging a window across the screen
Just open task manager and see if the CPU spikes.
Many thanks!

How can i restrict a user not open more than 3 sessions in database

how can i restrict a user not open more than 3 sessions in database.For this i have already create a profile and set CONCURRENT_SESSIONS=3 and assing that profile to user but the problem is there.

You will simply need to use the following parameter in your CREATE PROFILE syntax.
CREATE PROFILE normal_user
LIMIT SESSIONS_PER_USER 2;
but the resource limits set for a profile are enforced only when you enable resource limitation for the associated database.
you could do this either by having an initialization paramater which will invoke as soon as the database starts or try using ALTER SYSTEM command

High CPU usage when downloading Files

Hi there,
i noticed somethign strage when I download file off the internet or intranet. As soon as my download speed surpasses 1MB/s the CPU usage increases. To be exact the process kernel_task.
For example:
2,5 MB/s -> ~30%
10 MB/s -> ~60-70%
10+ MB/s -> ~80%
This is really annoying and I have now clue how to fix this. The Harddrive is a SAMSUNG HM320JI with a Journaled HFS+ partition. No Virus Scanner or anything. Also just reinstalled Snow Leopard yesterday.
Hope someone can give me some useful information how to fix this.

Hiya,
could it have something to do with the write speed/ cache/ram? On a "slow" disk, or one that is nearly full or one that is partitioned, it will take a great deal more ram and cache to "digest" the faster download and write it onto disk. I tend to do downloads to a separate drive. Have you got one? and can you repeat the high CPU usage when downloading / saving to the external drive? I note your drive might have 5,400 rpm, with "just" 8mB cache and 320 GB (quite a few SATA's have 7,200 rpm and 32 mB cache).
NB: If anything, I found SL more efficient, especially on my 4 year old Macbook, which is running much faster and has got more responsive after installation of (full version of) SL.
Message was edited by: Alexandre

OracleConnection opening more than one session.

Hi,
I am using ODP.NET 10.2 client, when i try to connect to Oracle database using OracleConnection object it is creating more than one session object.
I created one windows application, then i added the below code in form load event of the application
Please find the below code
string strConnection1 = "Persist Security Info=False;User ID=service1;Data Source=SD.World;Connection Lifetime=60;Max Pool Size=30;Min Pool Size=0;Pooling=true;PASSWORD=ases;";
string strConnection2 = "Persist Security Info=False;User ID=service1;Data Source=SD.World;Connection Lifetime=60;Max Pool Size=30;Min Pool Size=0;Pooling=true;PASSWORD=ASES;";
string strConnection3 = "Persist Security Info=False;User ID=service1_customer;Data Source=SD.World;Connection Lifetime=60;Max Pool Size=30;Min Pool Size=0;Pooling=true;PASSWORD=ases;";
OracleConnection con = new OracleConnection(strConnection1);
con.Open();
OracleConnection con1 = new OracleConnection(strConnection2);
con1.Open();
OracleConnection con2 = new OracleConnection(strConnection3);
con2.Open();
When i run the below Query
select * from V$session where PROGRAM='WindowsApplication7.vshost.exe'
I am getting 6 session. Please help me out in this regard.

The Error message was Unable to open connection once it reach the maximum session.
I agree for that, Previously I was not disposing that connection object. Presently I am using keyword to dispose the connection object.
Here I have doubt
In connection string the pooling attribute is true by default.
Scenario 1:
If I specify externally(through connection string it behaves very badly) like pooling=true in the connection string. It will create more than one session.
Scenario 2 :
If I donot specify pooling attribute in connection string it will create only one session and it will consume the same session for all the instance.
Below is the code snippet
Here is the example for scenario 2
private void button3_Click(object sender, EventArgs e)
//Connection string which donot have a pooling value.
string strConnection1 = "Persist Security Info=False;User ID= is_customer;Data Source=ISD.World;Connection Lifetime=60;Max Pool Size=30;Min Pool Size=0;PASSWORD=is;";
using(OracleConnection con = new OracleConnection(strConnection1))
con.Open();
MessageBox.Show(GetSID(con));
//con.Dispose();
using (OracleConnection con2 = new OracleConnection(strConnection1))
con2.Open();
MessageBox.Show(GetSID(con2));
//con2.Dispose();
Scenario 1 Example:
private void button7_Click(object sender, EventArgs e)
string strConnection1 = "Persist Security Info=False;User ID= is_customer;Data Source=ISD.World;Connection Lifetime=60;Max Pool Size=30;Min Pool Size=0;Pooling=true;PASSWORD=is;";
using (OracleConnection con = new OracleConnection(strConnection1))
con.Open();
MessageBox.Show(GetSID(con));
//con.Dispose();
using (OracleConnection con2 = new OracleConnection(strConnection1))
con2.Open();
MessageBox.Show(GetSID(con2));
//con2.Dispose();
GetSID is the function which is used to fetch the session of the connection string.
private string GetSID(OracleConnection con)
     OracleCommand cmd = new OracleCommand();
     cmd.Connection = con;
     cmd.CommandText = "select SYS_CONTEXT('USERENV','SID') from dual";
object sid = cmd.ExecuteScalar();
return Convert.ToString(sid);
In Scenario 2 example I will get only one session ID.
But in Scenarion 1 example I will get different session IDs.
Please suggest what is the problem.
Message was edited by:
user476285

Problem deploying more than one session beans(one is ok)

hi there,
im trying to deploy an ear that contains ejbs(entities and session). application goes ok if deploy it with any number of entities being called by a single session, i mean if there is only session bean in the ejb-jar.xml. it is deployed ok and works fine.
now when i try to add another session bean to the same ejb-jar.xml, it gives the following error during deployment.
The system is out of resources.
Consult the following stack trace for details.
java.lang.OutOfMemoryError
that's all what it prints and aborts deployment. again, its important that it deploys any number of entity beans problem is with more than one sessions.
thanx for any suggestions,
amt

Without seeing more details I don't understand why you would be running out of memory. We test applications with multiple session beans and we don't have that issue so I'm not sure why you would. Out of memory could be a bug in the deployment descriptor or the ear file or the JVM or the container or the deployment mechanism or maybe something else. That is the tough part of memory issue debugging at deployment time.
I would try to see if increasing memory on both the JVM for OC4J and for admin.jar (assuming you are using standalone OC4J when you deploy) to see if maybe it is something transient that is causing this to occur. Something else you could try is to use 1.3.1 JDK and see if the problem goes away. Beyond that I would try to get the simplest case you could that repeats the error and then let's take a look at it.
As for the exception hierarchy you describe, I don't see any particular problem with it at first look. Are you getting an error?
Thanks -- Jeff

Clamd high CPU useage

I have an Intel Xserve running 10.4.11
I followed Alex's tutorials some months back and upgraded clamav, amavisd and spamassassin.
All has been working well and I have been pleased with the performance.
However I decided approx 6 weeks ago to move from using clamscan to clamd.
On and off I have been getting exceptionally high CPU useage from clamd which requires me to kill off the clamd process as everything almost grinds to a halt.
I initially put it down to using Carbon Copy Cloner at night, but it has been getting more frequent during the day to the point where mail was just being queued.
I killed clamd again
ran
sudo /bin/launchctl unload /System/Library/LaunchDaemons/org.amavis.amavisd.plist
sudo /bin/launchctl load /System/Library/LaunchDaemons/org.amavis.amavisd.plist
but then all my incoming mail started going deferred
so I again reloaded amavisd.plist
then ran postsuper -r ALL, to requeue all the mail. This worked for a few minutes then clamd started going high CPU again.
Also in logs I started seeing in mail.log
Mar 25 12:30:50 fileserver postfix/smtp[147]: 502232589239: to=<[email protected]>, relay=127.0.0.1[127.0.0.1], delay=2616, status=deferred (host 127.0.0.1[127.0.0.1] said: 451 4.5.0 Error in processing, id=00791-05, virus_scan FAILED: virus_scan: ALL VIRUS SCANNERS FAILED: Clam Antivirus-clamd av-scanner FAILED: CODE(0x1801270) Too many retries to talk to /tmp/clamd (Can't connect to UNIX socket /tmp/clamd: Connection refused) at (eval 51) line 310. at (eval 51) line 511.; ClamAV-clamscan av-scanner FAILED: /usr/local/bin/clamscan DIED on signal 11 (000b) at (eval 51) line 511. (in reply to end of DATA command))
So I commented out the entry for using clamd in /etc/amavisd.conf
and rolled back to previous version of clamv.
I am now getting mail in, and although very brief high clamscan usage it is working, the queue has been cleared.
But now I am getting no logging in anything apart from one file, /var/log/asl.log.76
this has thousands of entries
[Time 2008.03.24 22:08:47 UTC] [Facility local6] [Sender clamd] [PID 18032] [Message accept() failed: Too many open files;] [Level 3] [UID -2] [GID -2] [Host fileserver]
syslogd is also running very high CPU.
Has anyone any ideas what caused my initial problems and what I may have done now to break the logging. ???

Thanks Alex
will schedule downtime to upgrade.
update
Reboot, again, of server cleared my logs problem (what a popular person I am )

3G High CPU utilization when waking

Recently gave my iPhone 3G to my wife, restored it to a new phone, then synced it with her information, music, applications, etc. She previously had used an iPod touch, and basically was trying to upgrade her to an iPhone.
The phone worked fine when I had it. I only had a short time with this specific phone as it was replaced under Apple Care about a week and a half before turning it over.
Now it pretty consistently is unresponsive when trying to wake it. By bringing up the System Activity Monitor App as soon as possible after powering it up, and leaving that app active when putting it to sleep I can see that the CPU gets slammed for a while during initial startup and has a significant period of very high CPU when waking. At this point I believe the cause of the unresponsive behavior is that the CPU is maxed out doing something.
I've compared the running processes to an iPhone 3Gs and they were identical except the 3G was missing the MobileMusicPlaye until I brought up the iPod on that phone.
There are 2 things about the initial setup of this phone for my wife that are causing me some concern;
1) When first connected to iTunes after the Factory Reset I was prompted to restore an older backup than the iPod Touch one I took right before the swap. I didn't realize it at the time and accepted that restore.
2) I was prompted to upgrade and followed through with an upgrade to 3.1.2 after the Factory Reset even though the phone was at 3.1.2 before the swap.
At this point I'm thinking about doing another Factory reset and set it up by sync only (no restore) but not looking forward to the time investment that will take.
Any ideas on where the high CPU utilization is coming from on wake? Is that a reasonable theory that the high CPU is causing it to be unresponsive when first woken up.
Any other insight from others that have been through this?

I ended up restoring the 3G phone to Factory settings and then sync'ing the desired information / applications onto it (did not restore the previous backup). The behavior is improved but this phone does seem to get bogged down when waking (observing high CPU utilization when monitoring with System Activity Monitor and experiencing sluggish / jerky response to user actions with no applications or just Mail running). The periods of non-responsiveness are not as long or as severe at this point, but still noticeable.
I'm at a loss since under the conditions I'm observing this there is nothing but Apple software running on the device (behavior is present with or without SAM showing me the CPU utilization). I'm a little suspicious that it might have something to do with re-establishing the WiFi connection, but it seems to continue for a bit after the WiFi icon appears in the status bar.
I'm probably going to make a genius appointment to see if they have any other ideas.

High CPU usage when downloading from ITunes Match

When attempting to download matched songs from ITunes Match, ITunes utilization spikes to over 70% and makes it almost impossible to do any other function on the computer. Eventually ITunes fails to respond. I have tried several remedies including reseting winsock. Is there any other options I should try?

Hiya,
could it have something to do with the write speed/ cache/ram? On a "slow" disk, or one that is nearly full or one that is partitioned, it will take a great deal more ram and cache to "digest" the faster download and write it onto disk. I tend to do downloads to a separate drive. Have you got one? and can you repeat the high CPU usage when downloading / saving to the external drive? I note your drive might have 5,400 rpm, with "just" 8mB cache and 320 GB (quite a few SATA's have 7,200 rpm and 32 mB cache).
NB: If anything, I found SL more efficient, especially on my 4 year old Macbook, which is running much faster and has got more responsive after installation of (full version of) SL.
Message was edited by: Alexandre

High CPU usage when updating or instlling s/w

When updating Win8 or installing MS Office system slows down and has to run all night to make any progress. Task Manager shows "System" using 49%. "go to process" shows "NT Kernel & System". The system was bought with Win7 and is now double booting both 7 & 8. Win 8 partition has 500 GB.
This question was solved.
View Solution.

Hello @Dan-44,
Welcome to the HP Forums, I hope you enjoy your experience! To help you get the most out of the HP Forums I would like to direct your attention to the HP Forums Guide First Time Here? Learn How to Post and More.
I have read your post on how your computer is displaying a high CPU usage when installing updates or other software. I would be happy to assist you in this matter!
What other programs are in use when these updates take place? How full is your computer's Local Disk hard drive? In the meantime, I recommend following the steps in this document on Resolving Slow System Performance (Windows 8). This should help improve the speed and efficiency of your operating system.
Please re-post with the results of your troubleshooting, as well as the requested information above. I look forward to your reply!
Regards
MechPilot
I work on behalf of HP
Please click “Accept as Solution ” if you feel my post solved your issue, it will help others find the solution.
Click the “Kudos, Thumbs Up" on the right to say “Thanks” for helping!

High CPU usage when editing large MXML files

I am using Flash Builder 4.5 Plug-in on OSX. When I have a large MXML file open, Eclipse will consume 100% of the CPU as long as that file is open, even if I am not doing anything. If I close that editor the CPU usage goes back down. I've seen this in Flash Builder standalone as well as the plugin. Is there a workaround to stop this CPU usage problem?
Thanks

Hiya,
could it have something to do with the write speed/ cache/ram? On a "slow" disk, or one that is nearly full or one that is partitioned, it will take a great deal more ram and cache to "digest" the faster download and write it onto disk. I tend to do downloads to a separate drive. Have you got one? and can you repeat the high CPU usage when downloading / saving to the external drive? I note your drive might have 5,400 rpm, with "just" 8mB cache and 320 GB (quite a few SATA's have 7,200 rpm and 32 mB cache).
NB: If anything, I found SL more efficient, especially on my 4 year old Macbook, which is running much faster and has got more responsive after installation of (full version of) SL.
Message was edited by: Alexandre

Multiple select queries used in Excel BI report ,fetching data from Sharepoint DB(SP2010_Prod_ProjectServer) causing blockage on DB ,when more than one workbook(same copy of Excel BI Report) refreshed using Refresh All option.

I am using mutiple select queries to fetch data from Project Server 2010 DB(its sharepoint DB) and these queries fetch data in Excel BI report by establishing connection with DB using instance name and all. I have enhance all these select queries and data
is being fetched in secs. but when more than one copy of same Excel BI report is refreshed using 'Refresh All' option, then these select queries cause blockage on DB.
Please let me know mitigation for this blockage issue.
Should I use begin transaction and commit transaction statements/ shared lock statements.
please reply

Hi,
run same query at the same time?

High CPU useage when more than 30 sessions logout/login

Similar Messages

Maybe you are looking for