Oracle utilizes only one core out of 8 core's.

Dear Team ,
We have noticed that, in oracle database server Oracle processes uses only one core and that goes to 100% usage remains are at idle possition thats the reason the Oracle performance is very poor some time system haults and needs to be started
Please suggest if you can have best solution,
waiting for your valueable response.
Regards,
Naveed Amir

860707 wrote:
We have noticed that, in oracle database server Oracle processes uses only one core and that goes to 100% usage remains are at idle possition thats the reason the Oracle performance is very poor some time system haults and needs to be startedNot unusual. It is often slower running processes across CPUs (e.g. 5 processes on 5 CPUs with 15% CPU utilisation per CPU), than running these processes on the same CPU (e.g. 5 processes on 1 CPU for a 75% CPU utilisation). See Non Uniformed Memory Access (NUMA) for some details on why this is the case.
The kernel deals with CPU scheduling. It decides which process/thread to execute on which CPU. From a s/w perspective, one can code "processor affinity" into your s/w to bind a process or thread to a specific CPU. Oracle only does this with Oracle XE - as this is one of limitation of the free version of Oracle's database s/w. Oracle uses default kernel scheduling for Standard Edition and Enterprise Edition versions.
Oracle uses the number of CPUs/cores to determine the processor load it can create. It does not force its processes to run on a single specific CPU (with XE being the exception).
So "blaming" Oracle for what you are seeing? You are blaming the wrong layer in the s/w stack. You are also very likely misinterpreting what you are seeing.
The #1 reason for poor Oracle performance is shoddy code. Code that you and your cohorts write and run in Oracle. And that is a far likelier candidate to blame for shoddy performance, than blaming Oracle and the kernel's process scheduling across CPUs.
And you have provided not a single shred of evidence that points to a SMP scheduling issue across multiple CPUs/cores.

Similar Messages

Parallelism in oracle using only one dual-core processor

Hi, is usefull or could i use parallel execution in oracle using only one dual core processor?
Can someone explain me why?

Hi
To take advantage of parallel processing you need enough resources (mainly CPU and I/O). If you run serially and the bottleneck is the CPU, then parallel processing might help. On the contrary, if the bottleneck is located in the I/O sub-system, throwing more CPU to it would not make things better. So, it's a question of balance between the CPU and I/O resources you have.
HTH
Chris

How can I deploy only one package out of a big project?

Does any one know how can I only deploy one package out of a big project?
We have a project which includes about 12 differenct packages. Is there a way in Jdeveloper for me to create a deploy profile to only deploy one package to a *.jar file?
Do I have to re-create a new project ( that's what I am doing currently) simply for deployment purpose?
By the way, click and pick class from more than 100 classes is too much of work. In addition, you don't really know exactly which class you are picking if two classes (in different packages) have the same name.
Thanks a lot.

Rename your LCA file extention into .ZIP
Open the ZIP file in any of the compressing utility (e.g. WinZip, WinRar etc.)
Extract the desired process and deploy it manually to your server.
Nith

[SOLVED] Only seeing 4 out of 8 cores (2x Quad Core)

Hello all,
I just got some new toys today. A server with 2 E5410's. 2 quad core 2.33GHz Xeons. Great fun. Now I have installed Arch64 on it. The problem is I can only see 4 cores, not 8. The BIOS shows 2 processors detected and 8 logical processors detected.
cat /proc/cpuinfo shows:
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 23
model name : Intel(R) Xeon(R) CPU E5410 @ 2.33GHz
stepping : 6
cpu MHz : 2333.416
cache size : 6144 KB
physical id : 0
siblings : 2
core id : 0
cpu cores : 2
fpu : yes
fpu_exception : yes
cpuid level : 10
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good pni monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr dca sse4_1 lahf_lm
bogomips : 4671.73
clflush size : 64
cache_alignment : 64
address sizes : 38 bits physical, 48 bits virtual
power management:
processor : 1
vendor_id : GenuineIntel
cpu family : 6
model : 23
model name : Intel(R) Xeon(R) CPU E5410 @ 2.33GHz
stepping : 6
cpu MHz : 2333.416
cache size : 6144 KB
physical id : 1
siblings : 2
core id : 0
cpu cores : 2
fpu : yes
fpu_exception : yes
cpuid level : 10
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good pni monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr dca sse4_1 lahf_lm
bogomips : 4668.31
clflush size : 64
cache_alignment : 64
address sizes : 38 bits physical, 48 bits virtual
power management:
processor : 2
vendor_id : GenuineIntel
cpu family : 6
model : 23
model name : Intel(R) Xeon(R) CPU E5410 @ 2.33GHz
stepping : 6
cpu MHz : 2333.416
cache size : 6144 KB
physical id : 0
siblings : 2
core id : 1
cpu cores : 2
fpu : yes
fpu_exception : yes
cpuid level : 10
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good pni monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr dca sse4_1 lahf_lm
bogomips : 4668.30
clflush size : 64
cache_alignment : 64
address sizes : 38 bits physical, 48 bits virtual
power management:
processor : 3
vendor_id : GenuineIntel
cpu family : 6
model : 23
model name : Intel(R) Xeon(R) CPU E5410 @ 2.33GHz
stepping : 6
cpu MHz : 2333.416
cache size : 6144 KB
physical id : 1
siblings : 2
core id : 1
cpu cores : 2
fpu : yes
fpu_exception : yes
cpuid level : 10
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good pni monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr dca sse4_1 lahf_lm
bogomips : 4668.31
clflush size : 64
cache_alignment : 64
address sizes : 38 bits physical, 48 bits virtual
power management:
Am I missing something?
Current kernel is 2.6.24.4-1
Will I need to compile the kernel myself with some different options? And if so, what options would those be?
Thanks in advance!
Last edited by DeltaXFX (2008-08-04 17:33:19)

I got it working and decided to wait it out until Arch officially supports 16 cores before I update the kernel again. These are the steps if I remember correctly...
Get the current kernel26 at http://archlinux.org/packages/core/x86_64/kernel26/
(you need to go to "view svn entries")
Download everything in there...
PKGBUILD
config
config.x86_64
kernel26.install
kernel26.preset
In "config" find the line "CONFIG_NR_CPUS" change it from "4" to "8" or whatever you want
"CONFIG_SPLIT_PTLOCK_CPUS" should also be changed to "8" if it isn't already
In "config.x86_64" look for the same two lines and make the appropriate changes.
Build and install the package as you normally would.
Restart, then cat /proc/cpuinfo should show a wonderfully long list of CPUs

Merge only one row out of multiple rows

I need to prepare a merge statement which merge only one row in to a table. Below is the table definition, sample data and merge statement i used.
      CREATE TABLE Mob_Attr (NAME VARCHAR2(20), ATTR_1 VARCHAR2(20), ATTR_2 VARCHAR2(20), ATTR_3 VARCHAR2(20));
      CREATE TABLE Mob_Desc (NAME VARCHAR2(50), ATTR_1 VARCHAR2(20), ID NUMBER(20));
      CREATE TABLE Mob_Main (NAME VARCHAR2(20), ATTR VARCHAR2(20), ID VARCHAR2(10));
      insert into Mob_Attr(name, attr_1, attr_2,attr_3) values ('n1','aa',null,null);
      insert into Mob_Attr(name, attr_1, attr_2,attr_3) values ('n2','bb',null,null);
      insert into Mob_Attr(name, attr_1, attr_2,attr_3) values ('n3','cc',null,null);
      insert into Mob_Attr(name, attr_1, attr_2,attr_3) values ('n4','dd',null,null);
      insert into Mob_Desc (NAME,ATTR_1,ID) values ('n1','aa',10);
      insert into Mob_Desc (NAME,ATTR_1,ID) values ('n1','aa',15);
      insert into Mob_Desc (NAME,ATTR_1,ID) values ('n1','aa',123);
      insert into Mob_Desc (NAME,ATTR_1,ID) values ('n1','aa',1455);
      insert into Mob_Desc (NAME,ATTR_1,ID) values ('n2','bb',455);
      insert into Mob_Desc (NAME,ATTR_1,ID) values ('n2','bb',135);
      insert into Mob_Desc (NAME,ATTR_1,ID) values ('n3','cc',15);
      insert into Mob_Desc (NAME,ATTR_1,ID) values ('n3','cc',52);
      insert into Mob_Desc (NAME,ATTR_1,ID) values ('n4','dd',1455);
      merge into Mob_Main ma using
      (select d.name, d.attr_1,d.id from Mob_Attr a, Mob_Desc d
      where a.name = d.name and a.attr_1 = d.attr_1
      ) q on (ma.name=q.name and ma.attr=q.attr_1)
      when not matched then
      insert (NAME,ATTR,ID)
      values (q.name,q.attr_1,q.ID);
      Above merge statement will actually merge entire data . But i want the first row will only get inserted to the base table (Mob_Main). If we take name 'n1' first row from table Mob_Desc with ID as '10' only should get inserted. I can go for a cursor to do this as i have huge data volume i prefer merge instead.

MERGE INTO Mob_Main ma
     USING (SELECT d.name, d.attr_1, d.id
              FROM Mob_Attr a, Mob_Desc d
             WHERE a.name = d.name AND a.attr_1 = d.attr_1 AND id = 10
            UNION
            SELECT d.name, d.attr_1, d.id
              FROM Mob_Attr a, Mob_Desc d
             WHERE a.name = d.name AND a.attr_1 = d.attr_1 AND id != 10
                   AND NOT EXISTS
                              (SELECT 1
                                 FROM Mob_Desc d1
                                WHERE     d.name = d1.name
                                      AND d.attr_1 = d1.attr_1
                                      AND d1.id = 10)) q
        ON (ma.name = q.name AND ma.attr = q.attr_1)
WHEN NOT MATCHED
THEN
   INSERT     (NAME, ATTR, ID)
       VALUES (q.name, q.attr_1, q.ID);
COMMIT;
select *
from Mob_Main
order by name, id;
NAME                 ATTR                 ID
n1                   aa                   10
n2                   bb                   135
n2                   bb                   455
n3                   cc                   15
n3                   cc                   52
n4                   dd                   1455
6 rows selected.

Only one audio out for a speaker set that requires 3 ??

Hi,
I just got my splendid MacBook pro today and I put my PC for sale simultaneously.
Switching to Mac was a charm - I'll save the details - but only one thing is still bugging me :
I've got a Logitech Z-640 speaker set. The center speaker is the one to connect to the computer (or audio card whatsoever).
Well, it has three plugs, since it's a surround speaker set.
But, my MacBook only has one (the earphone plug indeed).
So if I plug the "green" plug, everything seems to work.
I unpress the "matrix" button on the center speaker and they all work, including the subwoofer.
On my PC I used to have a small application that would "test" them (make each speaker play one by one, so you could tell that they are all controlled independently and therefore that surround sound is enabled).
I couldn't find anything like that on OSX - is there a feature like that ?
And do you think I'll still be getting surround sound effects even though I could only connect one of the three plugs ?
Any other suggestions ?
And is there some kind of devide, like a USB one, that would let me connect the three plugs ? Is it worth it ?
THANK YOU FOR YOUR ANSWERS !!!

Congratulations on your new mac. It is a great machine.
First of all, the MacBook Pro actually has 2 audio outputs -- they are just collocated. One is the analog headphone jack and the other is a mini toslink digital optical port.
The "green" plug you refer to is just a stereo analog input. The "matrix" feature of your speaker system is just simulating surround sound. If this is OK with you and you don't want or need true surround sound that is just fine.
There is equipment available that can separate and convert the digital optical output signal into discrete analog channels, but this equipment is highly specialized professional gear. You're not going to spend $1000 to hook up a $50 speaker set.
If you want true surround sound, you should upgrade to a speaker system that will accept a digital optical input, normally referred to as toslink or s/pdif.

The settings I need are the only ones locked out.

I am trying to have my settings allow third party apps but it won't work. I put in the admin password and I click the lock. Everything is unlocked except the one section of settings I need. You can find what I'm talking about in the General tab of Security & Privacy in System Preferences.

Is it grayed out? Have you tried control - clicking on the file and using the contextual menu?

With HP 8600 Plus, why must I replace ALL colors at same time when only one is out?

I got a message that my Yellow Ink was low, but when I replaced just that, the printer said I had to replace the other two colors as well. I thought the advantage of having them separate was that you could replace one at a time. Is there an easy fix? Something I am missing? Hope you can help.
Otherwise, am very happy with this printer.

When a cartridge is replaced there is an initialization routine to prime the cartridges. This will use some of the ink in the other cartridges. If they were near empty the initialization of the yellow cartridge may be enough to trigger the low ink message on the other cartridges.
There was a firmware update that addresses a few issues and includes inproved ink efficiency. See the page here for information and instructions to update the firmware.
Bob Headrick, HP Expert
I am not an employee of HP, I am a volunteer posting here on my own time.
If your problem is solved please click the "Accept as Solution" button ------------V
If my answer was helpful please click the "Thumbs Up" to say "Thank You"--V

Only one mac out of six has a speed problem

I have two Mac Minis, a MacBook, two Blue and White G3s, and a Gigabit G4.
Five of the computers return a 2.0 Mbps D/L speed when I run the ATT speedtest page for my DSL.
The MacBook returns a 375 Kbps DL.
Speedtests were run over a couple dozen times and speeds reported have been very consistent.
Why? Location doesn't seem to matter. Two Macs are in another room and the MacBook is in the same room as the wireless router and three other Macs. Yes, I moved the MacBook around and currently it's only 4 feet away from the router and within 2 feet of the three other wireless Macs.
All machines are running 10.4.8 with all updates applied. Essentially there should be no difference between the MacBook and the two Minis... similar processors and software.
I'm not a noob when it comes to the network settings and there is nothing setup different in the MacBook other than the static IP address of course.
What could be going on here to cause the DSL download speed of the MacBook to be only 16% of the other 5 computers?
Thank-you.

I think that a continuous, realtime monitoring of
available wireless networks could results in two kind
of problems: the software accessing the wifi card
resource, thus blocking access by the legitimate
manager software, and the network bandwidth being
occupied by continuous attempts of testing network
presence. The latter problem could partially hurt
also other connected computers, though it is divided
among 13 channels, only ne of which being really used
by your real network.
I agree.
This kind of problems is much more present in
Bluetooth, where the so called "aggressive polling"
is able to produce the equivalent of a denial of
service.
I also have Bluetooth turned on.
As long as just CoconutWiFi is killed, my speed is back to normal even with my Bluetooth.

PXE Failing on only one DP out of 85!?

I have one DP, on the other side of the world which wont PXE boot...
I've removed it all (Boot Images, Pxe, WDS) and reinstalled a few times... I think the issue is when I distribute the boot images. I just
see this in smsdpprov.log:
[1AEC][Sat 01/11/2014 06:05:09]:Content 'CAS0024B.28' for package 'CAS0024B' has been added to content library successfully
[1AEC][Sat 01/11/2014
06:05:09]:Expanding D:\SCCMContentLib\FileLib\6D02\6D02AFA1BBF5AA1CD36F350695B8DD8D5A27983F248EA59317C9119F3299190B
from package CAS0024B
[1AEC][Sat 01/11/2014 06:05:09]:Finding Wimgapi.Dll
[1AEC][Sat 01/11/2014
06:05:09]:MsiEnumRelatedProducts failed
[1AEC][Sat 01/11/2014 06:05:09]:FindProduct failed; 0x80070103
[1AEC][Sat 01/11/2014
06:05:09]:MsiEnumRelatedProducts failed
[1AEC][Sat 01/11/2014 06:05:09]:FindProduct failed; 0x80070103
[1AEC][Sat 01/11/2014
06:05:09]:Found C:\Windows\system32\wimgapi.dll
[1AEC][Sat 01/11/2014 06:05:09]:Expanding CAS0024B to D:\RemoteInstall\SMSImages
Any ideas?

SMSpxe.log and distmgr.log is very important for troubleshooting this case. Check the log first.
Juke Chou
TechNet Community Support

Background render stays on, but only one project out of many...

The % done window will sometimes jump back and forth very quickly and 3 sec. of video takes forever to load on timeline. Media is on a 4TB drive with 500 gbs left....all other projects work fine....Thanks

Turn off BG Rendering. Select the project file in the browser– go to File and click on Delete Project Render Files.
Russ

Why a table can have only one PK ?

Just wanted to know the reason why Oracle allows only one PK per table ?
I came across a situation where my existing large PK which is on a single (ID) column needs an alteration to add a new column (No).
ID column on which PK is based is now subjected to have duplication and that is the reason why i want to add a new column. The new column composition is not going to ensure uniqueness.
So, i need to create (alter) PK on two columns now.
It seems, in order to do that, i need to drop the existing PK and create a new one. Since this is a OLTP production database, business cannot afford to drop the existing pk.
When i tried to create a separate PK on two columns, Oracle gave me below err:
ORA-02260: table can have only one primary keySo, i wanted to understand why Oracle allows only one PK / table ? and considering above err, is there any way to make required changes to my PK without altering existing PK?
(Table is a large 30 GB partition table and PK is a non partitioned global index)
I think PROBABLY i need to drop existing PK and re-create the new one. But core intention for opening this thread is to know rational behind one pk/table concept.

OraDBA02 wrote:
Just wanted to know the reason why Oracle allows only one PK per table ?
It's not exactly Oracle's fault. The ISO SQL standard requires that a "PRIMARY KEY" constraint can only be used once per table. However, a UNIQUE constraint on non-nullable columns achieves effectively the same thing as a PRIMARY KEY. You can have multiple UNIQUE constraints.
Arguably the PRIMARY KEY constraint is unnecessary and could safely be dropped from the SQL language. In principle in the relational model a "primary" key is nothing more or less than a candidate key. All keys are equal in the relational model and there is no fundamental reason to designate a primary key at all. There is certainly no reason not to have multiple keys. In the SQL world though it is a very widely observed convention to use a PRIMARY KEY constraint for each table and to use it for all foreign key references.

Import one tablespace out of exp file

Hello,
I 'd would like to know how to import only one tablespace out of an export file. The import went wrong for one TS and I'd like to only retry that particular TS.
Tyia,
Cheers,
A.R

Tablespace clause is applicable only if you are working with transportable tablespaces and not otherwise.
The following keywords only apply to transportable tablespaces
TRANSPORT_TABLESPACE import transportable tablespace metadata (N)
TABLESPACES tablespaces to be transported into database
DATAFILES datafiles to be transported into database
TTS_OWNERS users that own data in the transportable tablespace set
Import terminated successfully without warnings.

AME is only using one processing core out of six.

I have not been able to figure out why AME is only using one of the six processing cores that I have in my Core i7 hex core processor. I'm running Windows 7 Ultimate 64 bit and the latest version of AME 2014 CC. When I render in Premiere Pro CC 2014 all cores are utilized. The codec that I'm using in AME is "YouTube SD 480P 29.97 - H.264". The video that is being encoded is fairly complex but it is only 30 minutes long and AME is estimating 18 hours to completion. That seems pretty excessive for that length of video.
Any help would be greatly appreciated.

This forum post may be related to your question.
Re: AME CC2014 Very Slow Encoding

Lightroom using only one core?

I am new to Lightroom, and a very ametuer photographer. I'm running 3.4.1 on a Win7 machine. I'm discovering that it is a very resource intensive program, and realizing that I may need to upgrade my computer to even utilize it. Before I do that though, I'm trying to figure out why Lightroom only seems to be accessing one processor on a dual-core machine. After I've had Lightroom open for a bit, when I click into a folder to view the photos in it, even when there are only 3 photos in the folder, it is taking up to 20 minutes to load the thumbnails. When I open task manager, one CPU is running at nearly 100% and the other one is barely being used. Lightroom basically shows up at 50% (though sometimes up to 53%). Any idea why this would be the case?
I've got a fairly small catalog (less than 10,000 photos) and I've got plenty of room on my hard drive. I may need more memory, but considering nothing else is running, Lightroom shouldn't be that desperate for space. I want to confirm that the 64bit version was installed, but I can't seem to find a way to confirm that.
Thanks for any thoughts.

I'd suggest a tool like Spotlight On Windows (it's free) that will show what's going on in your computer at a fairly detailed level with simple graphics.
http://www.quest.com/spotlight-on-windows/
And to look at the process in detail. Microsoft's Process Explorer (procexp.exe) from the SysInternals tools.
http://technet.microsoft.com/en-us/sysinternals/bb896653
Memory is certainly the first suspect, but I'm running fine on a machine with only 2G of RAM.
Over 248000 photos in my catalog, and the database is on a local disk while the images and XMP files are on a NAS.
My CPU is a quad core 2.5GHz Xeon, max cpu use is 44% and I get the grid of all images ever first page display in 20 seconds.
However I am using Windows XP and I suspect that Windows 7 might be struggling on your PC.
I daren't upgrade my machine to Win 7 because it works with XP - ain't broke, won't fix it
I did try quickly a 5000 image catalog on my netbook with 2G RAM and it was slow, but then the whole netbook is slow with Win7 and 2G whatever I do with it
By slow I mean about 2 minutes to load the initial grid view of all my images taken in 2011.

Oracle utilizes only one core out of 8 core's.

Similar Messages

Maybe you are looking for