User lock persists for killed process

Hi,
We have an app that acquires a user lock (using dbms_lock, release_on_commit FALSE).
We needed to kill a session but it remained in v$session (KILLED), even after killing the OS process. Fine, I know it will clear up eventually.
But the user lock (UL) persists. Does anyone know how it can be force-released?
Calling dbms_lock.release (using handle in v$lock) fails, even as user SYS, the error being that the user does not own the lock. But the owning session is killed so who can release it?
Eventually the session (and lock) disappers from v$session/v$lock but can the lock be removed before then, to avoid a DB bounce in an urgent case?
Thanks

If the process was truely KILLED and not MARKED FOR KILL then I do not know what could of happened. I cannot remember Oracle ever failing to release a user lock when the acquiring session was terminated. We use a fair number of UL; however, few of our processes have concurrent sessions that attempt to use the same UL. Mostly we just ensure that batch processes do not get submitted twice and that some on demand processes run single threaded or do not run if another customer has submitted the task.
HTH -- Mark D Powell --

Similar Messages

How could i get the kernel and user cpu usage for each process

Hi all,
In order to monitor the system CPU usage, I would like write a script to gather the kernel and user CPU usage for each process, like the prstat or top does. As always missing the shortlived kernel usage, prstat or top cann't get the precise CPU usage. I checked with the dtrace syscall, proc and fbt provider, but don't get which one is useful.
Please provide your comments and suggestion.
Thanks in adv

mail2sleepy wrote:
As I've studied the "dtrace" for a while, and seems Sun gives a pretty high score on this new feature.....I do want to know whether there's some probe can work for it, like writing a "dtrace" version prstat.You can write a prstat without dtrace. Because that's just polling at specific intervals and reading some process structures from /proc. You could have dtrace fire a probe every 5 seconds and read the same thing, but it wouldn't really be using any features of dtrace. Trhying to write it "in dtrace" doesn't make much sense.
What you could do that would be harder via other methods is to fire a probe at process exit that displayed the process information including total CPU time. They could print exactly when processes exited. Doing that without dtrace would be very difficult.
Darren

User not authorized for message processing

Hi,
I have a SM59 connection from SAP PI --> SAP R3.
Targethost sapr3T.mydomain.
Path prefix: /sap/xi/engine?type=entry
Process: From PI send a message to R3 (scenarie is a Server Proxy...)
User: PIPROXY has Role SAP_XI_APPL_SERV_USER
This is not enought I get error "User not authorized for message processing" when trying the connection from my PI server. (I tried with SAP_ALL and then it works)
What would be the right user role for this operation ?
Br,
Martin

Hi all,
I have the same problem in a XI -> BW scenario (Abap proxy).
I get error "User not authorized for message processing" when try to send message from PI to BW.
(I tried with SAP_ALL and then it works)
User: PI1REMOTE has Role SAP_XI_APPL_SERV_USER and SAP_XI_IS_SERV_USER
What would be the right user role for this operation ?
Thanks
Daniele

WF-BATCH user locks up inbound IDoc processing

Hi Experts!
We have an issue with our inbound IDoc processing randomly being locked by the WF-BATCH user. I have searched SAP Notes, and the only thing I could find was on Note 519420, it said to check the default settings for date format and decimal notation format in SU01. All looked fine, except for the decimal notation which was 1.234,567.89, instead of 1,234,567.89. If I changed this, would this fix the problem? Any suggestions would be greatly appreciated. Thank you in advance.
WC

WF-BATCH - is an ID in SAP which processes batch workflow tasks.
So if your IDOC posting triggers workflow task - it's going to be processed by WF-BATCH.

Mac user locks files for Windows users on NFS shares

Hi Everyone,
We run a mostly Windows environment here at my school and have two AD servers which also do our file shares for AD users. Every user has a network drive mapped to L on Windows, and this is mounted on the dock of our Mac users (all Macs are bound to the AD for auth and the OD for management). We also have shared folders for different groups of people, such as admissions or art (where the issue is). Most of these divisions are only Windows, so file sharing isn't and issue - but this is not the case in our Art department.
We have a shared folder (T on Windows) setup for users of the Art department with mostly Office docs in it. Windows users can write to the share just fine, just as the Mac users can. The issue comes up once a Mac user edits and saves the file (all to the share) - Windows users are now locked out of the file because the permissions are reset. What can we do to fix this issue? It's obnoxious having to go and reset the permissions after every edit from a Mac user.
Thanks,
-MRCUR

We also have that same problem. We don't actually solved the issue but we what we do did is just unlock the files once and even they are edited or opened in windows or mac, it does not lock the file. We use windows pc and connect to the drives and just simply right clicking the file then properties a button will show in the properties window saying "unlock". After we did that, those files was never locked out again.

Search for a process

If there are a lot of processes in the WorkSpace, is there a way for end user to search for a process by its name (in order to Start that Process)?
The WorkSpace help only talks about searching a process in the "Tracking" section. And I can't find a search box in the "Start Process" section.
thanks

Hi
We've found searching for processes is a very common requirement among our clients, and have developed an extension to Workspace to allow this.
It allows end user to search on any process type. Search fields are very simple to define, and you can control who has access to each process type.
You can see details here:
http://www.avoka.com/avoka/enquirytool.shtml
You can download a trial here:
http://www.avoka.com/apps/checkcookie?location=/apps/qpacdownload&qpac=y&qpac_code=Workspa ceEnquiryTool1.7
Howard
http://www.avoka.com.

Database for Completed Processes

Hi All,
in our business case, we need a possibility that user can search for completed processes.
I know about the Runtime View in GP, but there is a problem when the user, who is searching, wasn't attached to a processrole like Overseer, or Administrator.
Is there any solution?
Thanks for your help
Steve

I know GP Development has in their short-term plans to deliver analytical capabilities to GP and I would expect such requirement would be easily handle using Visual Composer

Scripts for killing blocking oracle processes?

We are having a problem in that when a user disconnects improperly, the oracle processes on our Oracle Apps box don't die properly and rows remain locked. For other terminal based versions, the script to kill the process is pretty simple, but for the web-based, java interface, it's unclear how I would go about doing this. Does anyone have scripts for doing this that could run via cron? The applications are running on solaris 2.6.
Thanks,
Nathan Olla
Unix Sys Admin

Take a look at note 130709.1 on Metalink. It contains configuration info for httpds.conf - your Apache configuration file. You can use this to time-out sessions properly. Hope this helps.

Custom exit or badi for changing the user fields in operation of process order

Hi all
    I want to find a exit or badi to set default value to the user fields in operation of process order when COR1.
    I have tried exit 'PPCO0001', but this exit is call after commit work, so I can not change any thing
    I also tried to use badi 'workorder_update', 'before_update', but all parameters in this BADI are importing, can't be changed, I've tried to change it and then program dump.
    Do you have any solution for this question?
Thanks

Do a Implicit Enhancement at the PBO of this subscreen.

BADI/User Exit for Order processing

Hi Experts,
I have a requirement that whenever someone modifies the order belonging to a specific order category then some error/ warning is issued. Can you please suggest some User Exit/BADI in APO for the same.

Hi Tiago,
My requirement is to block the processing of Released process order from both Product view and Detailed Scheduling Board.
If I am getting you correctly then in the mentioned BAdI and Method --I should call FM --/SAPAPO/OM_PEG_CAT_GET_ORDERS (to fetch orders by order category) to see whether the order category is for released process order or not.
If its Released process Order --then I should block the change. Is that correct for both Product View and Planning Board?

User Exit for Inbound Process - Changing IDOC control record

Hi,
I want to utilize the identification field (under details tab) of the IDOC Control Record. What is the suitable user exit for inbound process that i can use to change the IDOC control record?

It's been a long time since I have done any IDOC processing.
Go to SE37 and enter the function module IDOC_INPUT_ORDERS. Display the source code. Search for "CALL CUSTOMER". You should see several "CALL CUSTOMER-FUNCTION" lines. These are enhancements where you can manipulate the IDOC data. If you locate the correct one (where the data is read from the IDOC but not yet validated), you should be able to map the partner function value. Double-clicking on the number in quotes will bring you to an "EXIT_" function module. A "ZX" INCLUDE program is referenced. If you haven't used the exit yet, double-clicking on the ZX INCLUDE will create it. The values passed into the function module are available in the INCLUDE.
Ideally, you should create a project in transaction CMOD that references the chosen enhancement. This way you can activate/deactivate without commenting/uncommenting the code in the INCLUDE.

User Lock issue when processing IDOCS

Hi Folks,
We are pushing the data into SAP using IDOCS.During this process some IDOCS are getting failed due to User Lock on shipment header.Anyone here can share their experience in dealing with User Lock issues.
Thanks,
Kiran.

Kiran,
Only one can edit the document, it's standard fucntion. I think No note can resolve this. I can suggest you to run a back ground job very frequently ,which picks these status 51 IDOCs and reprocess. Report is :RBDMANi2
We can use Message class & error message number on selection screen, so that we can make sure that only those idocs which were failed due to Lock problem are taken by the job.
Reddy

Locks list for a user session

Hello,
Is there a function module or class/method which will give me the list of all locks of a user in the current session and not across all sessions (like transaction SM12). Function modules like ENQUE_READ and ENQUE_READ2 provides the lock list for a user across all sessions but not in the seesion in which the FM is being executed.
Please let me know if you have any information in this regard.
Thanks & Regards,
Saurabh

Hi saurabh,
Dont post duplicate thread . Its against the rule of engagement.
Locks list for a user session
If you have got the answer in the previous thraed then close this thread.
Regards,
Sujit

Claim lock for dead process in a trace file

Has anyone seen messages like this in a trace file?
Any clues on whats causing it?
claim lock for dead process][lp 0x39dc41d50][p 0x39e626f70.7242][hist x49514951]
[claim lock for dead process][lp 0x39de15308][p 0x39e626aa0.7179][hist
x49514951]
[claim lock for dead process][lp 0x39de15068][p 0x39e626aa0.7179][hist
x49514951]
[claim lock for dead process][lp 0x39dc41c00][p 0x39e626f70.7242][hist
x49514951]
[claim lock for dead process][lp 0x39de0f908][p 0x39e626aa0.7179][hist
x49514951]
*

BANNER
Oracle Database 10g Enterprise Edition Release 10.2.0.4.0 - 64bi
PL/SQL Release 10.2.0.4.0 - Production
CORE 10.2.0.4.0 Production
TNS for IBM/AIX RISC System/6000: Version 10.2.0.4.0 - Productio
NLSRTL Version 10.2.0.4.0 - Production

Group data locked error for MM01 using parallel processing

Hello gurus,
I am using Call txn method (MM01) Parallel Processing method ( around 9 threads ). The Materials are getting locked around 10 percent sometimes.
This is happening randomly ..one day i dont have any locking errors ..next day i have ...Any ideas why this could be..any prereq i need to check before executing the parallel processing..
Thank you in advance..
sasidhar p

Hi Sasidhar
I guess you are either extending the Sales Data or MRP Data. Just make sure that you are processing these transactions in a linear form for a single material. We can use parallel processing for different materials but for a single material if we go for parallel processing we can definetely expect the Lock Objects error.
Kind Regards
Eswar

User lock persists for killed process

Similar Messages

Maybe you are looking for