Trace sessions source

We run SAP-J2EE 6.4 SP 17 on linux with JVM 1.4.2 in a cluster environment with 2 nodes.
We found out that often, the system creates a few hundred sessions objects with the same login id within 1-2 minutes of time frame.
Each line (out of a few hundreds) contain the following attributes
User Name: [email protected]
Session ID : (J2EE1234567)ID0012345678DB12345678901234567890End
Created On:17:24 02/09/07
Last Access:17:24 02/09/07
Expiration 17:54 02/09/07
Application:<b>sap.com/irj</b>
Note that the application attribute does not provide any real information about the application and process that created the sessions.
Is there any way to trace those session.
Thanks

Hi,
You can change the settings of logs and traces in Visual administrator.
go to /usr/sap/SID/JCxx/j2ee/admin
Then,./go.sh
enter with administrator credentials.
Under SIDserver---sevices--Log configurator.
go to Locations tab at the right side.
Then go to com---sapengineservices--
trace.
Change the severity of the trace level as required there.
reward points if helpful..............

Similar Messages

Bug: Trace session from sessions report

Attempt to trace session against a 9.2.0.6 database, failed due to DBMS_MONITOR not available.
Is this a newer package that the database version needs to be verified and adjust to use this call instead for older databases? sys.dbms_system.set_boo_param_in_session

Actually what would be best would be to use this for < 10g databases:
dbms_system.set_ev (9,29,10046,X,'');
where level x would be in a dropdown box in the dialog when enabling tracing on the session.
As a developer and DBA, I enable trace at many different levels depending on what I am diagnosing.

How to trace session(3rd paty application) accessing oracle

Solaris 9
oracle 9i and 10g
I would like to know how to trace session which is 3rd party application currently running .
I need to collect statisctics for currently running session.
Thanks.
Message was edited by:
user539835

See this thread:
Re: SQL_TRACE help to newbie in oracle
Search the forum and you will also see how to manually decode most of the trace file.
Charles Hooper
IT Manager/Oracle DBA
K&M Machine-Fabricating, Inc.

Event Trace Session missing for Failover Cluster

I have a failover cluster setup and is managed via a Windows Domain controller for the failover cluster network. I am troubleshooting a potential issue with the failover cluster and the recommendation is to go into event viewer, app & service
logs, Microsoft, windows and look for failover cluster diagnostics and operations logs and these do not exist.
It appears they are created by having an event trace session associated with windows failover clustering but apparently it was't created when the cluster was created for some reason. I am wondering how to create the proper event trace session
in order to get these additional failover cluster logs created?

Hello,
the following forum mat be the better one for your needs:
https://social.technet.microsoft.com/Forums/windowsserver/en-US/home?forum=winserverClustering
Best regards
Meinolf Weber
MVP, MCP, MCTS
Microsoft MVP - Directory Services
My Blog: http://blogs.msmvps.com/MWeber
Disclaimer: This posting is provided AS IS with no warranties or guarantees and confers no rights.
Twitter:

Failed to start one or more trace sessions

Hi,
I just installed the analyser on a Win2012R2 Core server and can not get the analyser to start.
===========================================
Failed to start one or more trace sessions due to the followinf errors:
Error Details:
Unable to start ETW session MMA-ETW-Livecapture-d28fac22-61bc-4e52-ac03-60fa79d2ec62
Host Name: Localhost
Error in the application.
===========================================
This is with the provider:
Microsoft-Windows-NDIS-PacketCapture
I have this error on two core machines so I missing something?!
BR
Per

This happens if you are not logged in as administrator, or perhaps if a session is already running. Does it change if you run Message Analyzer as administrator? What do you see when you type "Get-NetEventSession" form PowerShell?
Paul

Client identifier to trace sessions

Dear all,
I have CRM application installed in oracle soa 10.1.3 , (OID-connection pool). At the DB level, Data source user name is displayed in v$session. I understand that user session can be traced using
client_identifier. I referred below links :
http://download.oracle.com/docs/cd/B19306_01/network.102/b14266/apdvprxy.htm#i1009003
http://download.oracle.com/docs/cd/B28359_01/network.111/b28531/app_context.htm#CIHCFDJG
I couldn't understand the above notes.. Does this client identifier to be set in DB or application ? how I can set trace for this OID users ?
I know this is a DOC question, But I cannot understand this as am a newbie.
Please guide?
Kai

The idea behind client identifier is that it should be set on the appserver so that session usage across pooled connections can be tracked.

Trace session based on CLIENT_IDENTIFIER

Hello all,
i am on 11.2.0.1....we have a sap application and trying to trace a session from DB end....and seems like the job that the end user runs jumps from one session to another....in this senerio...how would i trace a session ??? i read article about dbms_monitor package to trace it...but as i mention, sap keep chaning the session from one to another.....
i read the below article...
http://ocsurabaya.wordpress.com/2007/07/18/tracing-sql-in-oracle-database-10g/
in which it suggest to set the service name and module name....but the problem is its hard to get the module name as its doing quite a few things....it jumps from one module to another....is there anyway to trace a session based on its CLIENT_IDENTIFIER (from v$session)....as this looks like the domian user account for that user who is running the job....
as i mentioned during that time(when job is ran)...the sid changes, modeule changes...the only thing consistent is the CLIENT_IDENTIFIER ( which seems to be the domian user name )...
so how can i trace a session based on CLIENT_IDENTIFIER ??

user8363520 wrote:
Hello all,
i am on 11.2.0.1....we have a sap application and trying to trace a session from DB end....and seems like the job that the end user runs jumps from one session to another....in this senerio...how would i trace a session ??? i read article about dbms_monitor package to trace it...but as i mention, sap keep chaning the session from one to another.....
i read the below article...
http://ocsurabaya.wordpress.com/2007/07/18/tracing-sql-in-oracle-database-10g/
in which it suggest to set the service name and module name....but the problem is its hard to get the module name as its doing quite a few things....it jumps from one module to another....is there anyway to trace a session based on its CLIENT_IDENTIFIER (from v$session)....as this looks like the domian user account for that user who is running the job....
as i mentioned during that time(when job is ran)...the sid changes, modeule changes...the only thing consistent is the CLIENT_IDENTIFIER ( which seems to be the domian user name )...
so how can i trace a session based on CLIENT_IDENTIFIER ??is application 3-tier?
does application utilize connection pooling?

Sql trace session

hi all
i read some articals of how to trace oracle sessions and i have some question
i use database 10g
trace vs event ??? different between them ????
how to trace individual statement i hope to give me this point in ex:
due to i can't get the sql_id ,how to know the sql_is of specify statement??
thanks

861100 wrote:
hi all
i read some articals of how to trace oracle sessions and i have some question
i use database 10g
trace vs event ??? different between them ????
how to trace individual statement i hope to give me this point in ex:
due to i can't get the sql_id ,how to know the sql_is of specify statement??
thanksHow do I ask a question on the forums?
SQL and PL/SQL FAQ
below show how to trace single SQL statement
bcm@bcm-laptop:~$ sqlplus user1/user1
SQL*Plus: Release 11.2.0.1.0 Production on Tue Aug 28 07:08:09 2012
Copyright (c) 1982, 2009, Oracle. All rights reserved.
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
07:08:09 SQL> set autotrace trace explain
07:08:28 SQL> select sysdate from dual;
Execution Plan
Plan hash value: 1388734953
| Id | Operation      | Name | Rows     | Cost (%CPU)| Time     |
|   0 | SELECT STATEMENT |     |     1 |     2   (0)| 00:00:01 |
|   1 | FAST DUAL      |     |     1 |     2   (0)| 00:00:01 |
07:08:37 SQL>

Trace session

I have executed a SQL statement on two servers with exactly the same configuration,OS and DB version.I traced the sessions on both the servers executing the same SQL.Now i need to find out from those trace files which SQL is performing better then the other.what are the factors i need to cosider to come to a conclussion that this SQL's performance is better than the other SQL.The result of processed trace files using tkprof is as below.
//SQL in server A
SELECT SERVICE.TERM
FROM
abc.TARIF_SERVICES TS,abc.DIC_DATA SERVICE WHERE TS.N =:B2 AND TS.SERVICE=
SERVICE.CODE AND SERVICE.UP=9 AND SYSDATE BETWEEN SERVICE.FD AND SERVICE.TD
AND SERVICE.LANG=1 AND :B1 BETWEEN TS.FD AND TS.TD
call count cpu elapsed disk query current rows
Parse 1 0.00 0.00 0 0 0 0
Execute 5203 0.57 0.53 0 0 0 0
Fetch 5203 0.15 0.59 48 26071 0 5203
total 10407 0.73 1.13 48 26071 0 5203
Misses in library cache during parse: 0
Optimizer mode: CHOOSE
Parsing user id: 55 (recursive depth: 1)
Elapsed times include waiting on following events:
Event waited on Times Max. Wait Total Waited
---------------------------------------- Waited ---------- ------------
db file sequential read 48 0.01 0.33
//SQL in server B
SELECT SERVICE.TERM
FROM
abc.TARIF_SERVICES TS,abc.DIC_DATA SERVICE WHERE TS.N =:B2 AND TS.SERVICE=
SERVICE.CODE AND SERVICE.UP=9 AND SYSDATE BETWEEN SERVICE.FD AND SERVICE.TD
AND SERVICE.LANG=1 AND :B1 BETWEEN TS.FD AND TS.TD
call count cpu elapsed disk query current rows
Parse 1 0.00 0.00 0 0 0 0
Execute 5203 0.57 0.53 0 0 0 0
Fetch 5203 0.15 0.59 48 26071 0 5203
total 10407 0.73 1.13 48 26071 0 5203
Misses in library cache during parse: 0
Optimizer mode: CHOOSE
Parsing user id: 55 (recursive depth: 1)
Elapsed times include waiting on following events:
Event waited on Times Max. Wait Total Waited
---------------------------------------- Waited ---------- ------------
db file sequential read 48 0.01 0.33
Thank you all...

Formatting the output
SELECT SERVICE.TERM
FROM
abc.TARIF_SERVICES TS,abc.DIC_DATA SERVICE WHERE TS.N =:B2 AND TS.SERVICE=
SERVICE.CODE AND SERVICE.UP=9 AND SYSDATE BETWEEN SERVICE.FD AND SERVICE.TD
AND SERVICE.LANG=1 AND :B1 BETWEEN TS.FD AND TS.TD
call count cpu elapsed disk query current rows
Parse 1 0.00 0.00 0 0 0 0
Execute 5203 0.57 0.53 0 0 0 0
Fetch 5203 0.15 0.59 48 26071 0 5203
total 10407 0.73 1.13 48 26071 0 5203
Misses in library cache during parse: 0
Optimizer mode: CHOOSE
Parsing user id: 55 (recursive depth: 1)
Elapsed times include waiting on following events:
Event waited on Times Max. Wait Total Waited
---------------------------------------- Waited ---------- ------------
db file sequential read 48 0.01 0.33
//SQL in server B
SELECT SERVICE.TERM
FROM
abc.TARIF_SERVICES TS,abc.DIC_DATA SERVICE WHERE TS.N =:B2 AND TS.SERVICE=
SERVICE.CODE AND SERVICE.UP=9 AND SYSDATE BETWEEN SERVICE.FD AND SERVICE.TD
AND SERVICE.LANG=1 AND :B1 BETWEEN TS.FD AND TS.TD
call count cpu elapsed disk query current rows
Parse 1 0.00 0.00 0 0 0 0
Execute 5203 0.57 0.53 0 0 0 0
Fetch 5203 0.15 0.59 48 26071 0 5203
total 10407 0.73 1.13 48 26071 0 5203
Misses in library cache during parse: 0
Optimizer mode: CHOOSE
Parsing user id: 55 (recursive depth: 1)
Elapsed times include waiting on following events:
Event waited on Times Max. Wait Total Waited
---------------------------------------- Waited ---------- ------------
db file sequential read 48 0.01 0.33As Sybrand already mentioned,the results are identical.Is it somethig that is reported for either query that it is slow compared to the other one?If you suspect some performance degradation than you may want to trace both the queries with 10046 trace.Search this forum for this event and use it to trace the queries further.
Aman....

TRACE SESSION X TRACE USER (11gr2)

Hi,
We are facing performance isues on a Windows 2008 Oracle database (11.2.0.2) SAP BI Loads. There is a huge chain of jobs that are started and runs on the database with talkes a long time to execute. Let's supose that in this chain there are 500 jobs inside and takes more ou less than 15 to 20 hours. But, sometimes, this time increases and takes 30, 35, 40 hours.
So, we need to understand in which steps of the chain it takes more time, and one of the activities that we are planing is to trace all the execution and identify the waits.
The problem is:
. If there is a "father" process that call's the others, it would be easy since we could set the trace in this session.
Now,
How can we guarantee to trace all the chain since each step is one session opened on the database? Anybody know the usage or viability to trace the user at all? Any other options that could help us in this issue?
Doubt: Even that we set a trace on a "father" session, and this session calls/opens other sessions, this trace would contain all the instructions for all the sessions ou just the father? The trace sees a relation and traces all that are started in parallel?
Please disconsider tracing all the database. Not useful for this situation.
I aprreciate any Help

Hi,
this does not seem to be a security related issue. While the scheduler may know the notion of 'chains' the sql_trace facility
has no understanding of inheritance, so to gather trace info build it into the jobs so they can set event 10046 in their
respective job sessions. Also if you generate an AWR report for the duration of the jobs, you may be able to pinpoint
certain sql statements using more resources, compare with AWR report for the 'normal' situation,
greetings,
Harm ten Napel

How to trace the source of a ip packet?

On a router,how can I know the interface which a packet come in(I know the source ip address of the packet)?I use access-list and log to do it,but it's a slow method.Can I use debug or something to do it?
thank you!

You can use the 'debug ip packet' command. Please be careful with this command on heavly loaded routers tough.
The best practice would be to create a standard access-list with 'access-list 73 permit 10.0.0.1 0.0.0.0' or whatever ip address /subnet you want to track. Use this list in the debug command, 'debug ip packet 73' to narrow down the debug information....
/M

V$SESSION source table

Hi All,
what is base table for V$SESSION ?
And in which schema i can found it ?
Actually i want to search for a particular user login time and his PC information before 1 month
Regards

v$ tables are stored in control files.No they are not. They are in memory ONLY.
after successful startup mount command, you can
access to v$ tables & views.That part is mostly true. (Except that there are no 'V$' tables. The V$ view are actually public synonyms to V_$ views owned by SYS, which are built on X$ tables.)

Random Account Lockout (How to trace source?)

In Windows 2003 server native domain environment: XP Pro machines have no issues, but all ~10 PCs that have Win7 Pro (in different offices) have their domain accounts locked out randomly throughout the day. Workstations have no passwords listed in credentials
management.
Suspect it is something on the workstations that is sending incorrect logon and triggering the invalid password lockout limit on domain policy. Found MSFT tools to trace in XP, but nothing for Win7. Does anyone know how to use Procmon or similiar tool to
trace such source on the workstations? Thank you.
(Procmon.exe from systernals)

Hi,
The user account has been automatically locked because too many invalid logon attempts or password change attempts have been requested.
We can run the LockoutStatus.exe on domain controller to identify and investigate the account lockout issue.
Troubleshooting tools:
By using this tool, we can gather and displays information about the specified user account including the domain admin's account
from all the domain controllers in the domain. In addition, the tool displays the user's badPwdCount value on each domain controller. The domain controllers that have a badPwdCount value that reflects the bad password threshold setting for the domain are the
domain controllers that are involved in the lockout. These domain controllers always include the PDC emulator operations master.
You may download the tool from the link
Download Account Lockout Status (LockoutStatus.exe)
http://www.microsoft.com/downloads/details.aspx?familyid=D1A5ED1D-CD55-4829-A189-99515B0E90F7&displaylang=en
Once we confirm the problematic computer, we can perform further research to locate the root cause. Actually, there are many possible
causes for bad password, such as cached password, schedule task, mapped drives, services, etc. Please remove the previous password cache which may be used by some applications and therefore cause the account lockout problem.
Troubleshooting steps:
1. Click Start, click Run, type "control userpasswords2" (without the quotation marks), and then click OK.
2. Click the Advanced tab.
3. Click the "Manage Password" button.
4. Check to see if these domain account's passwords are cached. If so, remove them.
5. Check if the problem has been resolved now.
If there is any application or service is running as the problematic user account, please disable it and then check whether the problem
occurs.
For your convenience, I'd like to list the common troubleshooting steps and resolutions for account lockouts as the following:
Common Causes for Account Lockouts
To avoid false lockouts, please check each computer on which a lockout occurred for the following behaviors:
Programs:
Many programs cache credentials or keep active threads that retain the credentials after a user changes their password.
Service accounts:
Service account passwords are cached by the service control manager on member computers that use the account as well as domain controllers.
If you reset the password for a service account and you do not reset the password in the service control manager, account lockouts for the service account occur. This is because the computers that use this account typically retry logon authentication by using
the previous password. To determine whether this is occurring, look for a pattern in the Netlogon log files and in the event log files on member computers. You can then configure the service control manager to use the new password and avoid future account
lockouts.
Bad Password Threshold is set too low:
This is one of the most common misconfiguration issues. Many companies set the Bad Password Threshold registry value to a value lower
than the default value of 10. If you set this value too low, false lockouts occur when programs automatically retry passwords that are not valid. Microsoft recommends that you leave this value at its default value of 10. For more information, see "Choosing
Account Lockout Settings for Your Deployment" in this document.
User logging on to multiple computers:
A user may log onto multiple computers at one time. Programs that are running on those computers may access network resources with
the user credentials of that user who is currently logged on. If the user changes their password on one of the computers, programs that are running on the other computers may continue to use the original password. Because those programs authenticate when they
request access to network resources, the old password continues to be used and the users account becomes locked out. To ensure that this behavior does not occur, users should log off of all computers, change the password from a single location, and then log
off and back on.
Stored user names and passwords retain redundant credentials:
If any of the saved credentials are the same as the logon credential, you should delete those credentials. The credentials are redundant
because Windows tries the logon credentials when explicit credentials are not found. To delete logon credentials, use the Stored User Names and Passwords tool. For more information about Stored User Names and Passwords, see online help in Windows XP and the
Windows Server 2003 family.
Scheduled tasks:
Scheduled processes may be configured to using credentials that have expired.
Persistent drive mappings:
Persistent drives may have been established with credentials that subsequently expired. If the user types explicit credentials when
they try to connect to a share, the credential is not persistent unless it is explicitly saved by Stored User Names and Passwords. Every time that the user logs off the network, logs on to the network, or restarts the computer, the authentication attempt fails
when Windows attempts to restore the connection because there are no stored credentials. To avoid this behavior, configure net use so that is does not make persistent connections. To do this, at a command prompt, please type net use /persistent:no. Alternately,
to ensure current credentials are used for persistent drives, disconnect and reconnect the persistent drive.
Active Directory replication:
User properties must replicate between domain controllers to ensure that account lockout information is processed properly. You should
verify that proper Active Directory replication is occurring.
Disconnected Terminal Server sessions:
Disconnected Terminal Server sessions may be running a process that accesses network resources with outdated authentication information.
A disconnected session can have the same effect as a user with multiple interactive logons and cause account lockout by using the outdated credentials. The only difference between a disconnected session and a user who is logged onto multiple computers is that
the source of the lockout comes from a single computer that is running Terminal Services.
Service accounts:
By default, most computer services are configured to start in the security context of the Local System account. However, you can
manually configure a service to use a specific user account and password. If you configure a service to start with a specific user account and that accounts password is changed, the service logon property must be updated with the new password or that service
may lock out the account.
Internet Information Services:
By default, IIS uses a token-caching mechanism that locally caches user account authentication information. If lockouts are limited to users who try to gain access
to Exchange mailboxes through Outlook Web Access and IIS, you can resolve the lockout by resetting the IIS token cache. For more information, see "Mailbox Access via OWA Depends on IIS Token Cache" in the
Microsoft Knowledge Base.
MSN Messenger and Microsoft Outlook:
If a user changes their domain password through Microsoft Outlook and the computer is running MSN Messenger, the client may become locked out. To resolve this behavior,
see "MSN Messenger May Cause Domain Account Lockout After a Password Change" in the
Microsoft Knowledge Base.
For more information, please refer to the following link:
Troubleshooting Account Lockout
http://technet.microsoft.com/en-us/library/cc773155.aspx
Account Passwords and Policies in Windows Server 2003
http://technet.microsoft.com/en-us/library/cc783860.aspx
Hope this helps!
Novak

How to generate the trace files for remote db link session's?

User are complaining, the db link queries are performing slowness..
how to enable the sql trace session for db link's in remote database...
Is there any way to enable sqltrace for the dblink session ?
if not how to enable the sql trace for entire database level, rather than session based...

An explain plan of the SQL being ran on the local database will review the SQL being passed to the remote db. You can then explain that SQL on the remote db.
I have had to tune a few distribued queries so more than likely the explain plan alone will be enough to allow you to tune the query to improve performance. If not then you can go to the trouble of trying to set up dual traces.
HTH -- Mark D Powell --

Trace file created for the wrong session

DB version : 10.2.0.3.0
OS version: Solaris 5.10
2 node RAC
Using DBMS_MONITOR.SESSION_TRACE_ENABLE, I have been trying to trace an oracle session spawned by a C++ application.
So, i determined which instance this session is connected by querying gv$session. I logged in as SYS in that instance's Node and then issued
execute dbms_monitor.session_trace_enable(4371,98124, true, false);After enabling, i saw new A trace file being generated. But this trace is actually tracing SYS's session which executed dbms_monitor.session_trace_enable and not the session (4371,98124) !!! I tested few basic queries in SYS's session like
select sysdate from dualAnd i can see all these queries appearing in the trace file. What could be the cause of this issue?
I have noticed that
SHARED_SERVERS = 1 on both nodes. Is this causing this weird issue? For tracing, If i disable SHARED_SERVERS using
alter system set SHARED_SERVERS = 0
will it cause any connection problems for the client applications because of their tnsnames.ora file setting?

John_75 wrote:
Thank you DBMS_DIRECT.
trace is enabled only on the instance to which the caller is connected
So, how do i enable tracing for sessions spawned by Java or C++ applications.
According to the documentation, i can use DBMS_MONITOR for tracing applications
http://download.oracle.com/docs/cd/B19306_01/server.102/b14211/sqltrace.htm
Your 2 node RAC has two instances (e.g. INST1 and INST2). My understanding of that is that with the above command, to be able to trace session 123 connected to INST1, your SYS need to be connected to INST1. To be able to trace session 456 connected to INST2, your SYS need to be connected to INST2.

Trace sessions source

Similar Messages

Maybe you are looking for