Increase processes on AIX

Hi,
We are trying to increase oralce processes for one of our databases.
I'm looking for the list of parameters,which are affected
while increase the processes and what parameters need to be modified.
By the way the database is running on AIX 4
I really appreciate if you could help me in this.
Thanks
Stewart

One way to see which parameters are affected is to just try it! :)
Simple test: (N.B. on 9.2)
# export ORACLE_SID=testing2
# touch init_paramstest.ora
# sqlplus "/as sysdba"
Connected to an idle instance.
SQL> startup nomount pfile=init_paramstest.ora
ORACLE instance started.
Total System Global Area 97587500 bytes
Fixed Size 450860 bytes
Variable Size 46137344 bytes
Database Buffers 50331648 bytes
Redo Buffers 667648 bytes
SQL> set pagesize 0
SQL> spool parlist_before.txt
SQL> show parameters
SQL> spool off
SQL> shutdown
SQL> exit
# ( edit init_paramstest.ora and add processes=<your_new_value> )
Then startup instance again (nomount) and spool parlist_after.txt...
# diff parlist_after.txt parlist_before.txt
61c61
< dml_locks integer 1472
dml_locks integer 16463c63
< enqueue_resources integer 1692
enqueue_resources integer 384133c133
< max_rollback_segments integer 73
max_rollback_segments integer 30203c203
< processes integer 300
processes integer 30225,226c225,226
< sessions integer 335
< sga_max_size big integer 105976148
sessions integer 38
sga_max_size big integer 97587500250c250
< transactions integer 368
transactions integer 41Now this doesn't tell you what you need to adjust manually, e.g. for your application, and also AIX resource settings required. (one guess is that u need to adjust the umaxproc system parameter)
Was this neat or what? ;)

Similar Messages

Abnormal increment in background processes on AIX unix (oracle8i)

Hi
We have a problem running oracle8i on AIX unix. Our system is having two
parallel IBM RS6000 servers. Both running in HACMP env. We are using oracle8i with almost 350 clients.
For the last few days we observed that total no of oracle background processes abnormally increasing and touches 400 on each server. We got this information from following command
ps -ef |grep ora |wc -l
The moment it reaches 400 no more sessions are able to connect to the database.
One more thing we have observed that about 150 processes are defunct, which we get from using following commands
ps -ef |grep defunct | wc -l
On killing these processes using
kill -9 <pid>
system is unable to kill it.
When we query from v$session connecting to each servers we hardly found total number of sessions more than 225 on each server.
If anybody has any idea why so many processes are spawned and how to control them please mail me the solution. Or if u require some more information about the system and oracle parameters please let me know. My email_id is
[email protected]
Thanks
SHAKEEL AHMAD

You compare v$process with v$session.
If you have any defunc process, they wont appear in the v$session. But, they resides in the v$process. You need to kill them from the os prompt.
SELECT spid
FROM v$process
WHERE NOT EXISTS ( SELECT 1
FROM v$session
WHERE paddr = addr)
This query gives the osid of all the defunc process and kill them using kill -9 on os.
Are your connection coming from jdbc? Also, you can make use of DCD Dead Connection Detector.
http://asktom.oracle.com/pls/ask/f?p=4950:8:8284822833477458404::NO::F4950_P8_DISPLAYID,F4950_P8_CRITERIA:1784142376742
Jaffar
Message was edited by:
The Human Fly

Parallel_automatic_tuning increase processes

In my 8174 db (enterprise ed) on aix 51 with 4 cpu, i see processes parameter value equal to 192. But in my init , i set it to 20. I changed the value to 30 but show parameter/select from v$parameter it's always 192.
It's the case since i set parallel_automatic_tuning to true. When i set this parm to false, processes is what i expect.
Is it a normal behavior ?
Why process is set to 192 with my 4-cpu server ?

I found a formula but almost right :
processes will be set to the larger of :
- parallel_max_servers + 11 (bg process) + cpu_count x 4
- //max_servers x 1,2
and //max_servers should be equal to cpu x 2 x 5 = 40.
As my //max servers = 160, 160 x 1,2=192.
It's finally amost ok for me (with oracle formula, my //max servers should be = 40, not 160).

Find process in aix server

Hi all,
I want to "check for a process is running , if its running , go ahead and kill that else send an alert email"
, using shell script in aix server.
version:
uname -a
AIX XXXXXXXX
1 7 00CFC0F34C00
i googled online and got some ideas.
if [ "$(ps axu | grep process_name )" ]
then
kill statement goes here
else
email statement goes here.
Is this correct? or any other condition or modification do i need? Kindly advice.
Regards,
Pradeep. V

Moderator Comment:
You should go find an AIX discussion forum for your question.
AIX is not Linux and there might be idiosyncrasies in your OS that do not exist in other operating environments (such as Linux).
If you ask your question in an AIX forum you will get replies from people that know that OS.
Since your inquiry is beyond the scope of this LINUX discussion forum, it doesn't belong here.
It is now locked.

Top resource intensive processes on AIX/Unix/Linux

top is good to find this out, topas on aix, any other tips to find out memory/IO resource intensive processes, here i have for cpu....
oracle:tulpfsd01$ ps -e -o pcpu -o pid -o user -o args | sort -k 1 | tail -5r
%CPU PID USER COMMAND
1.1 1230914 oracle oraclepwpd (LOCAL=NO)
0.9 1248744 oracle oraclepwpd (LOCAL=NO)
0.9 1099146 oracle oraclepwpd (LOCAL=NO)
0.6 503688 oracle oraclepwpd (LOCAL=NO)
0.5 1239486 oracle oraclepwpd (LOCAL=NO)

Generally, memory shouldn't be a concern. The total amount of SGA is set (fixed) with SGA_MAX_SIZE and occupies shared memory (to be seen with 'ipcs')
The PGA is set by PGA_AGGREGATE_TARGET, but is only a rough pointer for oracle. Additional information about memory is seeable with pmap (alt least on linux and solaris)
You'll see the loaded shared libraries (which are used by the process in a shared matter, so not any process which uses libraries copy their own versions, and you should be able to identify the shared memory (identified with shmid=0xsomething) etc. On linux /proc/<pid>/maps shows the same in a little different way.
Now we've established that the shared libraries are used shared to preserve memory, and the shared memory is used shared (quite obviously), and 'put' into the process' address space, you should understand why listing VSS (virtual set size) and other memory stuff isn't very helpful: if you just add all listed used memory, there are much components which are not really used by that process.
Top and topas are helpful for finding CPU intensive processes. That is what both are designed for, to be able to identify the processes which are using the most slices of processor time. Of course, it's only helpful if you truly want to identify top CPU intensive processes. This is also quite easy to find inside the database.
IO: there's no way to identify processes which are the most intensive writers. Simple because linux doesn't keep statistics about it. Inside the database, it's easy to find (guess you should need an example of that?) please mind you should be aware how the database works: a user process does no writing, it's the logwriter, checkpointer and database writer who does that. The user process only reads.
I will see if I can find some nmon examples tomorrow, when I've got access to AIX systems.
If you still want examples, could you be specific?

Upgrade SQL server to increase processing speed

I have connected SQL server to my web service and currently the number of records being processed by a particular table has increased on a large scale. As a result of deleting and inserting into this table, the performance has gone down. Are there any
suggestions on how existing SQL server can be updated to increase the performance? Currently im using SQL server Enterprise edition 2012. Only a single server is been used with no replications. Please advice.
mayooran99

Hi
First of all, check out for fragmentation due to massive insert/delete operation an index is out of order
SELECT avg_fragmentation_in_percent FROM sys.dm_db_index_physical_stats
DB_ID('testdb'),
OBJECT_ID('dbo.T1'),
1,
NULL,
NULL
What does the above return?
To rebuild index/es
--single index
ALTER INDEX IX_Str ON tblname
REBUILD;
GO
--all indexes on the table
ALTER INDEX ALL ON tblname
REBUILD WITH (FILLFACTOR = 80, SORT_IN_TEMPDB = ON,
STATISTICS_NORECOMPUTE = ON);
GO
More over to speed up the query you need to have properly defined indexes on the table.
Best Regards,Uri Dimant SQL Server MVP,
http://sqlblog.com/blogs/uri_dimant/
MS SQL optimization: MS SQL Development and Optimization
MS SQL Consulting:
Large scale of database and data cleansing
Remote DBA Services:
Improves MS SQL Database Performance
SQL Server Integration Services:
Business Intelligence

Increasing Processing power by linking multiple pc's

Hi
Can Premiere Pro split its workload and distribute it across multiple computers that are connected via lan to decrease rendering time and speed up realtime viewing when working with muliple layers of video with lots of effects? Or is there a way to update the firmware on my xfx nvidia 8800 gt XXX edition graphics card to make it think its a quadro card?
Thanks
Russell

NO.
I have submitted several feature requests to add render farms. If you search for 'render farm' on the PR forums, you will find many posts on that topic.
Maybe it will help if you also submit a feature request. The more it is brought to the attention of Adobe, the better the chances of it getting implemented.

ORA-04030: out of process memory (PLS non-lib hp,pdz2M87_Allocate_Permanent

i am trying to validate a package body and getting the following error while doing that .
Any sollution for this
ORA-04030: out of process memory (PLS non-lib hp,pdz2M87_Allocate_Permanent

Hi Shakil,
Error: ORA 4030
Text: out of process memory when trying to allocate %s bytes (%s,%s)
Cause: Operating system process private memory has been exhausted
Action: See the database administrator or operating system administrator to
increase process memory quota. There may be a bug in the application
that causes excessive allocations of process memory space.
*** Important: The notes below are for experienced users - See Note:22080.1
Explanation:
The Oracle shadow process cannot allocate memory. The error
should show the number of bytes being requested and area where
they are required
Diagnosis:
1) Check for a subsequent error, Eg: ORA 7324, to see where
memory is being allocated.
2) Check user memory limits and swap space. If these are too
low you may get the above error.
Eg: 'limit' or 'ulimit -a' on Unix.
3) It may help to reduce <Parameter:SORT_AREA_SIZE> if this is large.
4) Check if using the Parallel Query Option - if an operation is
being performed in parallel the memory requirements are greater.
Best regards,
Rafi.
http://rafioracledba.blogspot.com/

What is the relationship between Process and Session parameter in Oracle

Hi All
I ran into a problem
"ORA-00020: maximum number of processes (150) exceeded " on the other day and I want to fix this temporary by increase the process by some number like 200.
Someone suggesting the syntax to fix my problem then the database must restart after running this command.
alter system set processes=200 scope=spfile;
However, I check the v$Paramenter view I see 2 values I believe they go together are
Processes: 150
Sessions:170.
Here are my questions:
1. "alter system set processes=200 scope=spfile;" is the correct way to increase the max Processes?
2. Do I have change the number of Sessions after changing the number of Processes?
3. Is there other way to increase the processes number without reboot the Oracle database?
Regards,
Jdang

According to the Oracle version# Reference manual which documents the database parameters sessions is set to 1.1 * process + 5 using the 10gR2 version.
Basically an Oracle process cooresponds to an OS process that performs the actual work for the session.
An Oracle session is a collection of memory structure entries that identifies and keeps track of what is being done for a client of the Oracle database.
Sometimes a single background process will support multiple sessions even in dedicated server mode such as when recursive SQL is performed to support the DML statement being done in the session.
PS - you should consider increasing processes first and only increase sessions if you still have an issue.
HTH -- Mark D Powell --
Message was edited by: mdp add PS
mpowel01

Formula to decide no.of process....

Hi All,
I am using ORACLE 11gR2. When I see the default value of processes and sessions is 150 and 247 respectively.
I want to know that is there any formula to decide the how many processes required like processes=1.1*session+5 etc.? And what is the optimal way to set both parameter.
In 10g I know that process and sessions having 50 number gap. But in 11g is more than this...
so Just for curiosity I want know if there is any formula to decide this.
Thanks...

Hello,
Login as sysdba
sqlplus "/as sysdba"
or
sqlplus '/as sysdba'
sql> show parameter sessions
sql> show parameter processes
sql> show parameter transactions
And if you are increasing sessions parameter you should consider increasing processes and transactions parameter as well.
Here is the formula you can use to determine their values.
processes=x
sessions=x*1.1+5
transactions=sessions*1.1
E.g.
processes=500
sessions=555
transactions=610
sql> alter system set processes=500 scope=both sid='*';
sql> alter system set sessions=555 scope=both sid='*';
sql> alter system set transactions=610 scope=both sid='*';
answered by OrionNet reference from the thread: Thread: How to increase Session
How to increase Session

Inactive sessions increasing database

Hi
We are using oracle11.5.10.2 on windows 2000 server and database version 10.2.0.3
we are facing problem with inactive sessions,grdually inactive sessions increasing and thats leads database crash.
Temporary i increased processes parameter value tp 400 prviously it was 200
Most of inactive session from apps user only
How can i fix this problem?
Thanks
With Regards
OH

Hi,
Please see these threads.
how to kill inactive sessions????????
how to kill inactive sessions????????
Inactive sessions in Database
Re: Inactive sessions in Database
Regards,
Hussein

Server Hangs while compiling some EJB in Weblogic 6.1 SP5 on AIX 4.3.3

Hi,
Iam not able to deploy a J2EE application in WebLogic 6.1 SP5 on an AIX 4.3.3
box.
CMP is being used and my EJBs are not being compiled( though there are no errors
thrown).
Iam using IBM DB2 ver 8.0 DB as client and server.
While compiling, the server Hangs, displaying the message "<Info> <Management>
<Configuration changes for domain saved to the repository.>"
Is there any configuration Issues with weblogic 6.1 SP5 and AIX 4.3.3?
Because the same EAR works in SP3.
Please help ASAP.
Thanks in advance.
Nagesh

The only thing to do again with "hangs" is to take thread dumps. As for why
things are going wrong...... I am at a loss other than some AIX weirdness.
Cheers
mbg
"Deshpande" <[email protected]> wrote in message
news:[email protected]...
>
Hi Mark,
I tried it out ....I ran EJBC on all my EJB's and then tried to deploy myapplication.
Application got deployed and the server was up, but JSP's dosent getcompiled(
i.e when I want to run my application, my server hangs again).
I tried one more option.. I pointed my JAVAHOME in startweblogic.sh to anSP3
JDK .. and it worked!! Now with this JDK Iam able to deploy my entireapplication
( with out precompiled) also run through the application without anyproblems.
I then over wrote the existing JDK ( of SP5 ) with that of SP3 and thentried.
Which also works fentastically!!!
I can feel something fishy here ....
Any help in this regard is most welcome.
Thanks and Regards...
Nagesh
"Mark Griffith" <[email protected]> wrote:
Deshpande:
"Deshpande" <[email protected]> wrote in message
news:[email protected]...
Hi Mark,
Thanks again.
For
Q1> How much time was between the two dumps?
A1> There was a minute gap B/w the two dumps. ( since it took so longto
write
the Core)
Q2> How long have you waited with no progress seeming to be made?
A1> Over a whole night. ( appx 14 hours).
Now I believe you would have figured out what the problem is.Not sure if this is a statement saying you think I have figured it out,
or a
statement saying you think I should have figured it out but I haven't.
Anyway... Did you try running ejbc on the "app" first?
Meanwhile also it would be great, if you would tell me the proceduresto
get support
form the BEA and AIX to solve the Issue.
http://www.bea.com/framework.jsp?CNT=index.htm&FP=/content/services/custome
r_support/contacts/
>>
I would suggest sending as much detail as possible, including zips of
the
thread dumps and pointers to this newsgroup conversation [email protected]
>>
I do however suggest you first try running ejbc.
Cheers
mbg
Thanks in advance.
R
Deshpande.
"Mark Griffith" <[email protected]> wrote:
Nagesh:
Wow, AXI outputs some interesting thread dumps. ;)
So the interesting thread (and the problem) is:
"main" (TID:0x300C41D8, sys_thread_t:0x30010A60, state:CW, native
ID:0x1)
prio=5
at java.lang.Object.wait(Native Method)
at java.lang.Object.wait(Object.java:429)
at java.lang.UNIXProcess.waitFor(UNIXProcess.java:130)
at weblogic.utils.Executable.exec(Executable.java(Compiled Code))
at weblogic.utils.Executable.exec(Executable.java(Inlined CompiledCode))
at
weblogic.utils.compiler.CompilerInvoker.compileMaybeExit(CompilerInvoker.j
a
v
a(Compiled Code))
at
weblogic.utils.compiler.CompilerInvoker.compile(CompilerInvoker.java(Inlin
e
d
Compiled Code))
at
weblogic.ejb20.ejbc.EJBCompiler.compileEJB(EJBCompiler.java(Compiled
Code))
at weblogic.ejb20.deployer.Deployer.runEJBC(Deployer.java(InlinedCompiled
Code))
at weblogic.ejb20.deployer.Deployer.compileEJB(Deployer.java(Compiled
Code))
at weblogic.ejb20.deployer.Deployer.deploy(Deployer.java(CompiledCode))
at weblogic.j2ee.EJBComponent.deploy(EJBComponent.java:33)
at weblogic.j2ee.Application.addComponent(Application.java:176)
at weblogic.j2ee.J2EEService.addDeployment(J2EEService.java:117)
at
weblogic.management.mbeans.custom.DeploymentTarget.addDeployment(Deploymen
t
T
arget.java(Compiled Code))
at
weblogic.management.mbeans.custom.DeploymentTarget.addDeployments(Deployme
n
t
Target.java:286)
at
weblogic.management.mbeans.custom.DeploymentTarget.updateServerDeployments
D
eploymentTarget.java:239)
at
weblogic.management.mbeans.custom.DeploymentTarget.updateDeployments(Deplo
y
m
entTarget.java:199)
at java.lang.reflect.Method.invoke(Native Method)
at
weblogic.management.internal.DynamicMBeanImpl.invokeLocally(DynamicMBeanIm
p
l
..java(Compiled Code))
at
weblogic.management.internal.DynamicMBeanImpl.invoke(DynamicMBeanImpl.java
C
ompiled Code))
at
weblogic.management.internal.ConfigurationMBeanImpl.invoke(ConfigurationMB
e
a
nImpl.java:374)
atcom.sun.management.jmx.MBeanServerImpl.invoke(MBeanServerImpl.java:1557)
atcom.sun.management.jmx.MBeanServerImpl.invoke(MBeanServerImpl.java:1525)
atweblogic.management.internal.MBeanProxy.invoke(MBeanProxy.java(Compiled
Code))
atweblogic.management.internal.MBeanProxy.invoke(MBeanProxy.java(Compiled
Code))
at $Proxy34.updateDeployments(Unknown Source)
at
weblogic.management.configuration.ServerMBean_CachingStub.updateDeployment
s
ServerMBean_CachingStub.java:3015)
at
weblogic.management.mbeans.custom.ApplicationManager.startConfigManager(Ap
p
l
icationManager.java:376)
at
weblogic.management.mbeans.custom.ApplicationManager.start(ApplicationMana
g
e
r.java:164)
at java.lang.reflect.Method.invoke(Native Method)
at
weblogic.management.internal.DynamicMBeanImpl.invokeLocally(DynamicMBeanIm
p
l
..java(Compiled Code))
at
weblogic.management.internal.DynamicMBeanImpl.invoke(DynamicMBeanImpl.java
C
ompiled Code))
at
weblogic.management.internal.ConfigurationMBeanImpl.invoke(ConfigurationMB
e
a
nImpl.java:374)
atcom.sun.management.jmx.MBeanServerImpl.invoke(MBeanServerImpl.java:1557)
atcom.sun.management.jmx.MBeanServerImpl.invoke(MBeanServerImpl.java:1525)
atweblogic.management.internal.MBeanProxy.invoke(MBeanProxy.java(Compiled
Code))
atweblogic.management.internal.MBeanProxy.invoke(MBeanProxy.java(Compiled
Code))
at $Proxy48.start(Unknown Source)
at
weblogic.management.configuration.ApplicationManagerMBean_CachingStub.star
t
ApplicationManagerMBean_CachingStub.java:480)
at weblogic.management.Admin.startApplicationManager(Admin.java:1309)
at weblogic.management.Admin.finish(Admin.java:650)
at weblogic.t3.srvr.T3Srvr.start(T3Srvr.java:547)
at weblogic.t3.srvr.T3Srvr.run(T3Srvr.java:214)
at weblogic.Server.main(Server.java:35)
So the Server never gets out of the Main thread to run, because theEJBC
compilation is "stuck", ejbc is running because either the EJB being
deployed was previously deployed on a different version of WLS orejbc
had
never been run on those beans. Either there is a problem on the AIX
VM
where execing processes are not returning, OR the ejb that ejbc isrunning
on is huge and is taking a REALLY long time. How much time was
between
the
two dumps? How long have you waited with no progress seeming to bemade?
Depending on the answers to these questions it maybe that there isa
VM bug
and you will need to contact BEA support to see the procedure forworking
with AIX to resolve the issue.
You maybe able to work around the problem by running EJBC on the ejb
in
question prior to starting the server.
http://edocs.bea.com/wls/docs61/ejb/EJB_utilities.html#1075296
Cheers
mbg
"deshpande" <[email protected]> wrote in message
news:[email protected]...
Hi, Mark,
Thanks for the early reply.
I have attached two consecutive dumps Zipped.
Pls look into the same and helpme out.
Thanks in advance.
Regards,
Nagesh
"Mark Griffith" <[email protected]> wrote:
When it hangs, you need to get thread dumps.
"kill -3 PID" is how this works on Solaris. Additionally if the
terminal/shell has focus that started the process you can do "CTRL
Break"
as
well. The thread dumps will go to std-out/err, so you'll have
to
redierect
std-out and err of the shell. (Make sure you don't nohup the
process
(if
AIX even supports that) since this prevents sigquits). When thehang
occurs
take a thread dump, wait 10 seconds and take another one, thenplease
zip
and attach the file in a post.
Cheers
mbg
"Deshpande" <[email protected]> wrote in message
news:[email protected]...
Hi,
Iam not able to deploy a J2EE application in WebLogic 6.1 SP5
on
an
AIX
4.3.3
box.
CMP is being used and my EJBs are not being compiled( though
there
are no
errors
thrown).
Iam using IBM DB2 ver 8.0 DB as client and server.
While compiling, the server Hangs, displaying the message "<Info><Management>
<Configuration changes for domain saved to the repository.>"
Is there any configuration Issues with weblogic 6.1 SP5 and AIX
4.3.3?
Because the same EAR works in SP3.
Please help ASAP.
Thanks in advance.
Nagesh

N- Level Process Controled Item workflow.

Hi Experts,
I am using Process Controlled item level workflow in SRM 7.0 (WS40000014) . My workflow is working for only one level approval.
I dont know how to increase number of approvers.... Could any one please suggest a BADI and process to achieve this?
Thanks in advance,
Regards,
Bhanu.

Hi Bhanu,
    Increasing the number of approval level( APPROVAL_INDEX ) is not same as SRM 5.0 in SRM 7.0.
     First you have to implement custom BADI for increasing Process level which is same (APPROVAL_INDEX), then implement another BADI to process the approval level. you can find the sample code for increasing the process level in standard Badi '/SAPSRM/BD_WF_PROCESS_SC'.
Best Regards,
Saravanan Dharmaraj

Processes Parameter in 11g

Hello,
Oracle offers licencing by the number of processes parameter. I want to increase Processes value to 300. let me know oracle recomends any licence aggrement? or we can increase
the processes parameter any value depends on the no of CPU's ? Is there any effect if we increase processes parameter to 300? FOR INCREASING THE PROCESSES PARAMETER ORACLE RECOMENDS LICENCE AGGREMENT?
NAME VALUE
processes 150
sessions 264
transactions 290
SESSIONS_MAX USERS_MAX CPU_COUNT_CURRENT
0 0 24
Above is the total cpu count.
Regards,
Edited by: Hari Ravipati on 24 Jun, 2012 4:36 AM

Oracle licensing options vary - pl check with your friendly Oracle salesperson regarding your license options. Any licensing advice offered in these forums in not legally binding in a court of law :-)
HTH
Srini

Tp import flag not properly set in "STMS", tp-processe not ended

since last Stopping of our Q-System the display in of the import queue doesnt work fine. when we transport requests from the development-system to the q-system the truck-icon doesn't disappear. the processes in AIX are not finished, after time the "overview of the transport logs" shows that the import is done, but the truck doesn't disaappear
we tested the transport tool (RSTPTEST): o.k., in the TMS configuration we made the communication test: o.k.
sometime we can make the truck disappear, when we delete the .LOB in /usr/sap/trans/tmp but this doesn't work with all requests

Check the Import monitor and the tp system log. If the truck icon still shows is most likely because the Import is not fully finished, As I said check the import monitor and let us know what it says.
Regards
Juan

Increase processes on AIX

Similar Messages

Maybe you are looking for