Looping processes in unix

Hi,
Can anyone tell me how do i check the looping processes in unix and how can i resolve it.

I am not sure what you trying to ask, but my general understanding of you question tells, that you are trying to find out a session with lots of CPU utilitzation(looping)
on your command prompt type:
oracle(FC4/) % prstat
or
oracle(FC4/) % top
PID USERNAME SIZE RSS STATE PRI NICE TIME CPU PROCESS/NLWP
25291 oracle 3288M 3251M sleep 21 0 0:00.26 2.2% oracle/2
by looking at under the "PID" column and co-relating the value with "CPU" column value will tell you which session/process are consuming lots of CPU, as you have mentioned already that there are some looping processes.
hope it helps.
Cheers,
Z
Message was edited by:
Zahid Lakhani

Similar Messages

Processes in UNIX

Hi guys.
I'm trying to execute a process in Compaq UNIX Tru64, but I can't do it!
The code I use is the following:
import java.io.*;
public class RunProcess {
public RunProcess(String[] args) {
if (args.length == 0) {
System.out.println("You missed command");
System.exit(0);
try {
Runtime rt = Runtime.getRuntime();
Process p = rt.exec(args);
int i = Integer.MIN_VALUE;
while ((i = p.waitFor()) != 0) {
System.out.println("waiting...");
System.out.println("Finish" + p.exitValue());
catch (Exception ex) {
ex.printStackTrace();
public static void main(String[] args) {
new RunProcess(args);
If I run this code in Windows, for instance using:
java RunProcess WinWord MyDocument.doc
I've no problem: the MS-Word document appears on my screen and the Java program loop within while until I close the MS-Word document. After that the Java program prints "Finish" and the session in stopped.
If I trying to execute another program (not java!!!!!!) on a UNIX PC using the same RunProcess class, the Java program always loop within while (it continuously prints "waiting...") and it never stops.
Anyone of you could explain such different behaviour from Windows to UNIX?
Thanx.

You should NOT loop in waitFor! It is sufficient to call it only once!
public abstract int waitFor() throws InterruptedException
Waits for the subprocess to complete. This method returns immediately if the subprocess has already terminated. If the subprocess has not yet terminated, the calling thread will be blocked until the subprocess exits.

Launch background process from unix command line and detect if it is runnin

in Cocoa, I am able to run unix app by using NSTask, it works fine to me. But I want to do the following:
1. Start up a background process.
2. Wait to see if this background process has been launched correctly.
3. If yes, keep running following codes...
With NSTask, I cannot make sure when the background process has been launched correctly... and other codes just keep running...
If I do something like:
while (bgprocess isRunning)
}

What exactly do you mean by "launched correctly"? Do you just care that the process has begun? Do you want to make sure it is still running? Do you want to see if there is some signal emitted (either affirmative "I am running" or absence of a child-died signal)?
Your isRunning example does appear to be in accord with the NSTask docs for "do blah blah blah while the task is running", but that doesn't seem to be what your description says you want to do. Is "blah blah blah" an idle loop until the task has started, or is it something to do while the task is actually running?
How about looking up the pid of the child process (processIdentifier for the NSTask) and then check whether that process is still running?

How to design a loop process in BPEL

Dear all:
Thanks in advance for your any suggestion.
Scenario
I try to implement a BPEL which runs some activities periodically until meets some condition.
Problem
1. How should I goto the start activity of the loop when the process finishes one cycle?
2. Could BPEL terminate in the middle of the cycle rather than in the end of the cycle?
Pls apologize the unclear description, waiting for your words.
BR,
Robert

I dont understand the exact requirement, but you may try
1)desing bpel process A and bpel process B. have your repeatative tasks as a part of process B. Keep Wait Activity in process A to halt process execution. After the wait will be over, the process B will again gets executed. You may like to use While loop withing process A instead of creating process B.
2) You can think about scheduling BPEL process. the server will make sure to execute the BPEL process at specific time intervals.

How can I see the memory of a process in Unix

I want to see how much memory a process with a specific pid takes in Solaris. How can I do this? I don�t have the top command available.
Thank�s in advance

use the ps command with the -vp switches then the pid (i.e. ps -vp 123)
to show the percentage of memory the process is taking
(note: Top shows the same info that this shows, I have a feeling that
you want to know the exact amount of memory in bytes that the process is
taking. In that case you should post this question in a Linux/Unix/Solaris
forum for better answers)
Hope that helps.

Process Scheduler UNIX - PSDSTSRV no post

Hello I have a trouble.
I have defined a Process Scheduler in UNIX… the processes and reports finish correct, but the distribution finishes incorrect… in the file stdout I have the following message:
OprId = VP1
PS_TOKEN=AAAAnwECAwQAAQAAAAACvAAAAAAAAAAsAARTaGRyAgBOZwgAOAAuADEAMBQa2m1pVz/gsTCheZFtS1f62cS47wAAAF8ABVNkYXRhU3icHYpLDkBAAEPfILN0ExPzQRwAO5mE2Fq6pMOpaZPXpinwmKpuMEjV+9NykfGWhZ2tVT9YObk15ESgl2c6MRSmQo8janEMSqfH78jEyAcNZwp8
The XML file returned by the web server is invalid. (63,94)
XML document object creation failed. (63,102)
Unable to process HTTP reply from Report Repository. (63,73)
=================================Error===============================
Message: Unable to post report/log file for Process Instance: 7075, Report Id: 5359
Process Name: AEMINITEST, Type: Application Engine
Description: Programa Simple Pruebas AE
Directory: /psoft/pt849/appserv/prcs/fs9dmo/log_output/AE_AEMINITEST_7075
=====================================================================
Can anybody help me?

Its hard to tell why report distribution fails , there is a lot of criteria that needs to be in place to make it work.
Does the user have process scheduler admin/report distribution admin roles assigned ?
Is local node setup with user id/password ?It could be unix report node definition pointing to an incorrect location.
what is the reports location in web server config.properties file ?

Auditting Java Process under Unix

Is there a way to see which java processes are running
in Unix?
ps -f does not work because the arguments passed
to java are too long and get truncated

Hi,
Try this command:
# top | grep java
"top" utility you can download it from http://www.sunfreeware.com for appropriate solaris versions.
Revert back.
Thanks,
Senthilkumar

Problem loop processing with a foreground task

Another problem: I have taken the example task provided by SAP to create a sales order in the foregrouund by passing it data from an ORDERS IDoc (the task is TS20000615). This task calls transaction VA01 with the data from the IDoc converted into a BDC table, and then displays what is happening as a foreground process.
I wanted to have this in a LOOP until a sales order nuumber was created; my thinking was if somebody starts the task and is unable to save the order at that time, the existance of the order number would be the LOOP condition, and a new work item would appear in the recipients inbox.
When I tried this, I end up in a constant screen processing loop, never seeming to exit the task, i.e. it looks as if when I cancel the processing for one task, it automatically starts the next and displays the screen processing.
Is there an obvious mistake here, or am I doing something that is fundamentally flawed?

Anthony,
What you need is task based on an asynchronous method with a terminating event. This will keep the same WI active until it is succesfully completed instead of generating a new WI for each execution.
Change the property of your BDC method to Asynchronous.
Define a terminating event for your task (inputFinished?).
Cheers,
Ramki.

Killing the process at unix level

Hi friends,
I wanted to kill a process at the unixlevel, can any one help me.
In SMLT i have scheduled a language it is running from last two days, i tried to check in SM66,SM50 and SM04 there is nothing there.
I have checked in SM37 also there is no job there.
But in SMLT i can see the tranport icon.
I want to kill from the unix level . Pl help.
thanks.

There is no easy way of "deleting that icon" - various tables are involved in the language import, they apparently contain the wrong state.
I suggest you either debug the program yourself and "fix" that manually or create an OSS call and let support people have a look on the system.
Markus

Identifying lingering PeopleSoft process on Unix

I am applying at PeopleTools patch. The doc says to shut down any appservers, web servers, process schedluers, and EM Agents before applying and make sure there are no lingering threads from any processes or things could get messed up. I shut down the appserver and process scheduler and the EM Agent on unix. The webserver is running on Windows along with another appserver and process scheduler.
However, when I make a backup of the PS_HOME before patching, the backup copy is a slightly different size, suggesting that a file is not copied because it is in use. The difference is the backup copy has a smaller appserv directory and a slightly larger verity directory.
How do I make sure all processes are killed and also get a complete backup?
I am doing the copy as root and getting subdirectories and keeping the same permissions. There are other PS_HOMES on this server.

Yes, on unix, my backup copy of the PS_HOME for that environment is about 300K smaller than the original.
Sessions to the database don't show anything from that PeopleSoft user/home owner.
i dont want the patch to fail (due to something in use) then the backup be invalid. Our unix admin did some work with the fuser command and revealed nothing. He thinks I shouldn't worry about the slight difference.
I did notice something strange though after a while, a .adm folder appeared in the backup home. It contained only what appeared to be a license key.

LOOP PROCESSING

Hi,
I have one query regarding loop.
suppose i have fields like f1,f2,f3,..f5.
now all the data's are stored in wa.
is it possible to make the loop like
loop at wa-f1 to wa-f5.
processing.
endloop.
Is it process for wa-f1, wa-f2, wa-f3, wa-f4, wa-f5.
If it is not possible then pls suggest me.

Salil,
If I understand you right, your fields are different fields with different names within the same structure like this:
data: begin of wa,
f1 type p,
f2 type p,
f3 type p,
f12 type p,
end of wa.
If this is so, then LOOP is not a way to attack your problem. Instead have a look at this construct:
DO 12 TIMES VARYING f FROM wa-f1 NEXT wa-f2.
processing. Here just sum up the values
tot_f = tot_f + f.
ENDDO.
You have to make sure all fn fields are of same type and are separated from each other with the same distance.

Process in Unix

How to bring process to background.?
How to bring process to foreground.?
How to deburg an shell script file.?
How to find errors in shell scripting file.?

How does your questions relate to the Oracle Forms product?
Craig...

HOWTO: start/stop htmldb processes in Unix

I've installed 10gR2 DB, and then installed HTMLDB from the companion CD. I could access the server no problems, so my configuration is good and working.
The issue is how do I start/stop the processes. It appears (from this forum) that Windows installs create a service but that's of course no good to me as I'm on AIX.
I'm a database guy, not an iAS guy, and it took me a while to figure out the way to make this all work.
First, set your ORACLE_HOME to the HTTP home (not the DB home).
In ORACLE_HOME/opmn/bin is the magic command:
./opmnctl startall
or
./opmnctl stopall
Seems to do the job.

That's almost right - that command will start & stop the Oracle HTTP Server (Apache). If Apache is down, then there is no way to access HTML DB via a web browser. However, users - if they had access - could still access the database, of course.
There is no such thing as the "HTML DB Process", as HTML DB is nothing more than a schema in the database.
Thanks,
- Scott -

Multiple processes on unix server even stop replicat, is that block port?

I have configured on source server Oracle goldengate manager/extract on port 7809 and same configured on target server with port 7809.
Now over the weekend, I encountered two problems.
1). tcp ip timeout from extract error log and extract abend.
test : telnet linuxserver 7809
$ telnet linuxserver 7809
Trying...
Connected to linuxserver.
Escape character is '^]'.
Connection closed.
It connected, then after a second, connection closed.
2). On target server, there are multiple processes for replicat existed even I stopped replicat and stopped manager. There are nothing running.
ps -ef|grep ggs
oracle 13595 1 0 May20 ? 00:00:00 ./replicat INITIALDATALOAD -p 7819-12818 -m 7809 PARAMFILE /u01/app/oracle/product/11.2.0/ggs/dirprm/load2.prm REPORTFILE /u01/app/oracle/product/11.2.0/ggs/dirrpt/LOAD2.rpt PROCESSID LOAD2 USESUBDIRS
oracle 14031 1 0 May20 ? 00:00:00 ./replicat INITIALDATALOAD -p 7819-12818 -m 7809 PARAMFILE /u01/app/oracle/product/11.2.0/ggs/dirprm/load2.prm REPORTFILE /u01/app/oracle/product/11.2.0/ggs/dirrpt/LOAD2.rpt PROCESSID LOAD2 USESUBDIRS
oracle 14447 1 0 May20 ? 00:00:00 ./replicat INITIALDATALOAD -p 7819-12818 -m 7809 PARAMFILE /u01/app/oracle/product/11.2.0/ggs/dirprm/load2.prm REPORTFILE /u01/app/oracle/product/11.2.0/ggs/dirrpt/LOAD2.rpt PROCESSID LOAD2 USESUBDIRS
oracle 14674 1335 0 13:34 pts/0 00:00:00 grep ggs
oracle 14962 1 0 May20 ? 00:00:00 ./replicat INITIALDATALOAD -p 7819-12818 -m 7809 PARAMFILE /u01/app/oracle/product/11.2.0/ggs/dirprm/load2.prm REPORTFILE /u01/app/oracle/product/11.2.0/ggs/dirrpt/LOAD2.rpt PROCESSID LOAD2 USESUBDIRS
oracle 15529 1 0 May20 ? 00:00:00 ./replicat INITIALDATALOAD -p 7819-12818 -m 7809 PARAMFILE /u01/app/oracle/product/11.2.0/ggs/dirprm/load2.prm REPORTFILE /u01/app/oracle/product/11.2.0/ggs/dirrpt/LOAD2.rpt PROCESSID LOAD2 USESUBDIRS
oracle 28295 1 0 May20 ? 00:00:00 ./replicat INITIALDATALOAD -p 7819-12818 -m 7809 PARAMFILE /u01/app/oracle/product/11.2.0/ggs/dirprm/load2.prm REPORTFILE /u01/app/oracle/product/11.2.0/ggs/dirrpt/LOAD2.rpt PROCESSID LOAD2 USESUBDIRS
oracle 28751 1 0 May20 ? 00:00:00 ./replicat INITIALDATALOAD -p 7819-12818 -m 7809 PARAMFILE /u01/app/oracle/product/11.2.0/ggs/dirprm/load2.prm REPORTFILE /u01/app/oracle/product/11.2.0/ggs/dirrpt/LOAD2.rpt PROCESSID LOAD2 USESUBDIRS
oracle 29158 1 0 May20 ? 00:00:00 ./replicat INITIALDATALOAD -p 7819-12818 -m 7809 PARAMFILE /u01/app/oracle/product/11.2.0/ggs/dirprm/load2.prm REPORTFILE /u01/app/oracle/product/11.2.0/ggs/dirrpt/LOAD2.rpt PROCESSID LOAD2 USESUBDIRS
I have to manually kill those processes by kill -9 on server.
How to control this manner?
thanks.

I have another quetion, in real life migration, our database have a lot clobs datatype, and logsegments, etc. What is the best practice or advices you can give to me?
I am probably not the best guy to answer this question.
Can I create multiple extracts on source and multiple replicat on source?
Yes.
Let's say we have 10 application users in the prod database, do I create at least extracts for each schema? then within each schema, I create multiple extracts to based on datatype or others?
Yes. You can create one Extract/Replicat group per schema. That would help in parallelizing the things.

O/S oracle process in Unix

How can I get the sid in v$session if I know the Oracle process on the O/S side? Thanks in advance for the reply.

Connected to:
Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
SQL> select process from v$session where username = 'DBADMIN';
PROCESS
9534
SQL> !ps -ef | grep -i 9534
bcm       9534 1553 0 19:00 pts/0    00:00:00 sqlplus
oracle    9535 9534 0 19:00 ?        00:00:00 oraclev112 (DESCRIPTION=(LOCAL=YES)(ADDRESS=(PROTOCOL=beq)))
bcm      10085 9534 0 20:17 pts/0    00:00:00 /bin/bash -c ps -ef | grep -i 9534
bcm      10087 10085 0 20:17 pts/0    00:00:00 grep -i 9534Edited by: sb92075 on May 10, 2010 8:18 PM

Looping processes in unix

Similar Messages

Maybe you are looking for