Major problem - Process Flows hanging randomly - 11.2.0.1

Hi,
We are experiencing the following problem.
Randomly process flows sit doing nothing.
Have ran some days and totally fine. Process flows fine - each activity has all the necesassry outgoing activities.
What causes this?
If run again - fine.
How do we resolve.
We need to be able to kick off the process and leave it to run without having to check if it hanging al the time.
Any help/advice would be great.
Many Thanks

Hi,
I found out that wrong ordering of transitions may cause strange problems. In owb, check that the outgoing transistions of each operator are numbered sequenty (0,1,2, ...). Sometimes, some number appear twice, e.g. 0,1,1). This often happens after an existing transition was deleted.
And make sure that the "else" transition (the one without any condition) always has the highest number.
Regards,
Carsten.

Similar Messages

Process flow hangs in state "bussy"

Hi,
We are encountering what seems to be a "random" problem when runing process flows. In about 1 out of 4 runs the flow "hangs" with status "busy"
The technical specs are as follows:
OWB 9.2.0.2.8 on WIN2K Server SP4.
There are two runtime repositories on two different WIN2K database servers (Oracle 9.2.0.4 EE)
The symptoms are as follows:
- We have more than 10 process flows containing from 15 to 60 processes. The problem is not related any single process flow, and occurs on both our target servers.
- The process flows runs OK in approximately 3 out of four runs.
- When the process flows "hangs", one or more of the processes in the process flow also hangs in the "busy" state. The hanging process can be a mapping or an external process
- The processes that have status "busy" have actually finished. They have a completion time set, and in the case where the process sends mail, the mail has been received. In the case of mappings I have verified that the mapping has completed, and that the session performing mapping has terminated.
- When the problem occurs, there is always the same "lock" pattern displaying in OEM:
Username Session IdLockMode HelMode Requ Object Name ROWID Object OwnObject T Resource ID1 Resource ID2
OWBRT92 17TM ROW SHARNONE WB_RT_SERVICE_QUEUE_TAB OWBRT92 TABLE 61161 0
OWBRT92 17UL EXCLUSIVNONE 1073741824 0
OWBRT92 48TM ROW SHARNONE WB_RT_NOTIFY_QUEUE_TAB OWBRT92 TABLE 61145 0
OWBRT92 67TM ROW EXCLNONE WB_RT_NOTIFICATION_COMPONENTS OWBRT92 TABLE 61137 0
OWBRT92 67TM ROW EXCLNONE WB_RT_NOTIFY_QUEUE_TAB OWBRT92 TABLE 61145 0
OWBRT92 67TM ROW EXCLNONE WB_RT_NOTIFICATION_ACKS OWBRT92 TABLE 61141 0
OWBRT92 67TM ROW SHARNONE WB_RT_NOTIFICATION_STREAMS OWBRT92 TABLE 61135 0
OWBRT92 67TM ROW EXCLNONE WB_RT_AUDIT_EXECUTIONS OWBRT92 TABLE 61199 0
OWBRT92 67TX EXCLUSIVNONE 262151 87564
Session 17 is the runtime service, and these two locks are present under normal operation. Session 48 is the Sqlplus session runing the execution template, and session 67 is (i think) the session that is causing the problem.
- The advanced queue WB_RT_NOTIFY_QUEUE tab shows "Number in state ready" equal to the number of processes still in the state "busy" i the process flow.
- When the runtime service is stoped (using the stop_service.sql) The locking sessions disappears.
- When the service is restarted, the processes change status to "bussy recovery" There is now a RPE-01008 and a RPE-01009 message in the "errors" section for the mapping.
- After the failed recovery, everything looks normal, except that there are now twice the number of messages in state READY on the WB_RT_NOTIFICATION_QUEUE. The only way I have found to get rid of these is to delete them from the queue table.
Is this a known problem? We are about to go live in a few weeks, and this could be a serious show stopper if we can't rely on the daily process flows to run as expected.
I could log a TAR on this, but I suspect support would ask for a test case, and Im going to have a hard time providing it for them in this case.

Jean-Pierre,
Thanks for the update. I was afraid this was going to be one of those "hard to pinpoint" issues...
When the problems occurs the Worflow Monitor always shows the task status as "Notified".
The "Main routine" is no longer executing. The session that was executing the mapping has disconnected, and the mapping has completed successfully. (I have verified this by checking the rows in the target tables). Also, the problem is not restricted to mappings. Sometimes it happens to "mail notifications", and in these cases the mail message is received as expected.
It looks to me like there is some kind of synchronization problem between AQ and the controlling process (the runtime access user session) that can arise under a given condition (which I unfortunately have not been able to determine what is). As far as I can tell, the process executing the mapping completes and puts the message that it has completed on the Notify_Queue (hence the messages in stat "Ready" on the queue), but for some reason the process waiting to read the message fails to do so.
As for a test case, I have recently sent an export of our design repository to Igor Machin. Is it possible for you to have a look at that? As I mentioned in the first post, our Process Flows are fairly complex. Some of the flows are executing 50+ mappings in parallel.

Process Flow hangs

I'm executing a process flow with several activities.
After each activity I check for status and error out if the activity fails. If the activity succeeds I go on to the next activity.
When I execute a map within the process flow, the process flow executes successfully only if the map is the last activity. The map functions to output data to a file.
The Process Flow hangs if after the execution of the map
a stored procedure is called. The stored procedure dynamically creates a DBLINK and updates a table.
When I cancel the execution of the process flow I see the map has successfully outputted all
records. But why hasn't it gotten to the stored procedure.
Each activity on its own works fine but in combination with each other they hang the process flow.

Mahesh,
I am assuming that RAB stands for Runtime Audit Browser.
The Runtime Audit Browser tells me how many records were
rejected but not which ones and why.
The debugger does not give adequate information in fact it does not issue a "break" where it should.
Are there any tools OWB provides for "catching", or
auditting data that cannot for whatever reason be extracted to a file.
I see that they have a mechanism for capturing problems with data coming from a file into a table. I don't see
where there is an explanation of error handling going from a table into a file.
Thanks!
Marion

Process flow hangs when mapping execution fails

Hi,
Has anyone faced the 'process flow hanging' problem after upgrading to OWB 9.2.0.4? Don't know why this is happening again after the upgrade. The audit browser clearly indicates one of the mappings has failed and workflow tables indicate there was a notification upon failure. The process flow runs several sub process flows each containing 5-6 mappings.
Also, the mapping that failed (due to unique key violations on the target table) executed the post mapping stored procedure although the post mapping run condition was set to 'ON SUCCESS'. Is this the expected behaviour for the post mapping operator when there are database related failures?
Our environment is OWB 9.2.0.4 on AIX 5.2.
Database version is 9.2.0.5. We recently upgraded the database from 9.2.0.4 to 9.2.0.5 and OWB from 9.2.0.2.8 to 9.2.0.4. Process flows are executed using sqlplus.
Thanks,
Chethan

Hi,
I would expect you to have less issues with 9204 as there were some fixes applied. So I have not seen this before.
To get into the mapping a little bit more, is this a row based mapping that is under the tresshold? In any case, if you specify "on success" I would expect it to only work on success... A database related error will also cause it to get an error, so this is interesting as well.
Does the error repeat at any case if there is a failure?
Jean-Pierre

Problem with program hanging randomly on certain commands from win 7 pro client to SB Server

Having a problem with program hanging randomly on certain commands from Win-7 Pro Client to SB Server Both 64-Bit
Five other slower XP-Pro 32 Bit systems though they are older and slower systems do not seem to hang as readily if at all.
It has been very frustrating as the Client-System, SB-Server and Program should work w/o any hitches, but this seems to work fine @ times and then hang randomly.
Would appreciate any and all suggestions in assisting.... JimBAgde-MSS

You can try this, as I have had similar problems with another MS Access .MDB file and slow access before. This fixed my problem for me. On the slow computer, make sure the program is set to see the .mdb file via UNC path, not a mapped drive letter. ex.
USE: \\yourserver\shared\dental\file.mdb
DO NOT: S:\\shared\dental\file.mdb
hope this helps

Process Flow hangs on wb_rt_notify_queue_tab

Hi,
We are encountering some problem on process flow.
Environment: Oracle 9.2.0.5, OWB 10g, workflow 2.6.2, HP-UX
The process flow has about 20 mappings. When it hangs, I can see the following:
1. In Workflow, the status of the mappings are "NODIFIED"
2. In wb_rt_notify_queue_tab, the DEQ_TIME shows null and the state is 0
3. In wb_rt_audit_executions, the return_result shows "OK" for this mapping, and elapse shows the seconds
it spent. and the data was loaded correctly.
4. There are some locks showing
SID 31 has locks on WB_RT_AUDIT_EXECUTIONS, WB_RT_NOTIFICATION_ACKS, WB_RT_NOTIFICATION_COMPONENTS, WB_RT_NOTIFICATION_STREAMS, WB_RT_NOTIFY_QUEQE_TAB, WB_RT_SERVICE_QUEQE_TAB.
SID 25 has locks on WB_RT_SERVICE_QUEQE_TAB
SID 97 has locks on WB_RT_NOTIFY_QUEQE_TAB, this is the SQLPLUS session invoking the Process Flow
I saw some messages on the similar topic in this forum, but I guess that's for early version of OWB.
Any suggestion will be appreciated
Thanks in advance.

I am also having same problem and my environment is
Environment: Oracle 9.2.0.3, OWB 10g, workflow 2.6.2, sun solaris 2.8 - 64 bits. Does installing/upgrading OWF 2.6.2 to 2.6.3 solve the problem and does it have any ripple effect on the system if you upgrade the OWF from 2.6.2 to 2.6.3? Please help.....

Process flow hanging in owb 11.2.0.2

Hello All,
My process flow keeps on running and never completes but in the design center I see that my maps have all completed, I checked my tables and the data has been loaded.
There is this process running in the background. I don't know what this is trying to do.
How could i solve this problem.
select /*+ FIRST_ROWS(1) */ tab.rowid, tab.msgid, tab.corrid, tab.priority, tab.delay, tab.expiration ,tab.retry_count, tab.exception_qschema, tab.exception_queue, tab.chain_no, tab.local_order_no, tab.enq_time, tab.time_manager_info, tab.state, tab.enq_tid, tab.step_no, tab.sender_name, tab.sender_address, tab.sender_protocol, tab.dequeue_msgid, tab.user_prop, tab.user_data
from "OWBSYS"."WB_RT_NOTIFY_QUEUE_TAB" tab
where q_name = :1 and corrid = :2 and (state = :3 ) for update skip locked
SELECT MIN(MSGID)
FROM WB_RT_NOTIFY_QUEUE_TAB N1
WHERE PRIORITY = (
SELECT MIN(N2.MSG_PRIORITY)
FROM AQ$WB_RT_NOTIFY_QUEUE_TAB N2
WHERE N2.CORR_ID = :B1 AND MSG_STATE = 'READY') AND N1.CORRID = :B1
BEGIN wb_rt_notification.wait( :1 , :2 ,
:3 , :4 , :5 )
; END;

Hi,
We are dealing with the same issue. Is this solved yet at your site?
I also note there are lots of records in WB_RT_NOTIFY_QUEUE_TAB. A lot of the AQ$_WB_RT_NOTIFY_QUEUE_TAB_E records are pretty old
Can I just delete them?
SQL> select count(*), Q_NAME from owbsys.WB_RT_NOTIFY_QUEUE_TAB group by Q_NAME;
COUNT(*) Q_NAME
80223 WB_RT_NOTIFY_QUEUE
751129 AQ$_WB_RT_NOTIFY_QUEUE_TAB_E
best regards, Hans

Owb Process Flow getting hanged/stuck .

Hi,
Problem description :- We have designed a Process Flow in OWB our project. The Process Flow is scheduled to run at 12.00 AM daily Dutch timings.
We have been facing a problem for around 3 months wherein the process gets hanged/stuck in between without giving any specific error message.
There isn't any kind of locking happening in the background as well. The OWB log gives the message "Forcibly deactivating execution with audit ID 1660235 ",
During the time at which the process hangs The process gets hanged in different timings in different sub processes(not in a particular process, so no specific pattern has been observed.)
To complete the process flow, we need to kill the process and rerun it again. When reran, it runs perfectly without any error.
The above error message appears in the log only when a process hangs. There is no such kind of logging during a successful execution.
Can you please let us know when and why does this message appear in the OWB log file.
Is there any specific kind of settings or configuration parameters that needs to be set to resolve this.
Regards,
Rosy Mukherjee

You might want to check to see if you have anything "parallel" running
We recently had an issue with one of our mappings where one day it decided to just stop working and "freeze"
Our DBA found that we had a Oracle hint for "Parallel" in the code and removed it. After that the mapping ran fine.
Best of luck

When exporting many videos with Lightroom 5, the process hangs randomly and needs to be restarted.

When exporting few videos, the process hangs randomly. The processor performance is 0 % and the lightroom says it is working (actually it is not). I have to quit lightroom, kill all processes like amecommand.exe, Adobe Media Core, Adobe Dynamic Link Manager and restart the process. It stops randomly.
I am exporting ALL-I videos from Canon 5D mark III, into H264, high quality.
PS: The same situation was with Lightroom 4.

Hello,
i have the same problem on LR5.5 on Win 7 (8G RAM)
Thank you for any help.
Petr

I am experiencing some major problems with my MacBook Pro. I have had some issues with it turning on/off at random times, but today, when starting, I get the grey start-up screen and a recovery bar. After filling in approx 1/4 of the way, the machine dies

I am experiencing some major problems with my MacBook Pro. I have had some issues with it turning on/off at random times, but today, when starting, I get the grey start-up screen and a recovery bar. After filling in approx 1/4 of the way, the machine dies. After starting it in recovery mode, it will not allow me to download OS X Mavericks- it says the disk is locked. Any ideas? I do not have a back-up and do not want to erase everything before I have explored my options. Help?

try forcing internet recover, hold 3 keys - command, option, r - you should see a spinning globe
most people will tell you to do both pram and smc resets (google) and if you still have issues, either clean install (easy) or troubleshoot (hard)

Process flow execute "complete" in Runtime Audit Browser but it hangs

I create a process flow that contain 4 sub-process flow, each sub-process flow contains several mapping. Each mapping in the process flow will be executed one by one if previous mapping execute successfully. If any error or warning, it will be ended with "End_Error" or "End_Warning". And I deploy the process flow to the workflow server successfully.
Sometimes the process will hang in some certain mapping when execute the process flow. After long time waiting, the status of process flow execution in Audit Browser will be "Complete", but in fact, the sub-porcess ended with that mapping and the next mappings will not be execute. This can be monitored in the Audit browser, as if the whole process "break" at that mapping.
In workflow monitor, I found the workflow has not complete yet. The status of that mapping node is "Notified", not "complete". So I have to abort the workflow in the monitor.
Sometimes when I try again it will be ok. If I deploy the process to another instance that in dw instance, it will also be ok. The dw instance use owb recommend parameters when we install db instance. All the mapping use the default parameters except "Default operation mode", here we use "row based".
Can somebody give me some hint?
Thanks.
OWB 10g rel 1, db: 10g rel2, workflow: 2.6.4

Hi Naidu
Yes you are absolutely right. When you press the cancel button in the GUI, the runtime repository does not get the signal that he is been asked to stop and abort. The GUI cancels himself but the background oracle process is still on !!!. The only way to do is to kill the oracle session. Another way I tried to tackle is
go to the directiry OWB/10.1/owb/rtp/sql and run the file @list_request.sql in sqlplus. Note that you have to login as runtime repositiry owner. After you get the audit id, run the file abort_exec_request.sql and supply the audit id that you obtained by running list_requests.sql. Sometimes it does not work also . Hence the best method is to kill the process

User Defined Activity "hangs" Process Flow if it runs for more than 10mins

I have been using OWB (10g r2) for a while and have used "User Defined" activities in production process flows successfully before.
However my problem now is that the activity (shell command) I want to perform, takes more than 10 minutes to complete. If I run the process flow with a parameter that generates a shell command that executes in two or three minutes, the process flow continues past the sucessfully executed user defined activity to the next activity.
BUT
If I run the same process flow (without modification or even redeployment) with a different parameter , the user defined acivity sucessfully executes (after about 11 minutes or so) but the process flow , doesnt seem to know that it is finished and just says it is still executing.
my "user defined" activity has
COMMAND: /usr/bin/ksh
PARAMETER_LIST bound to a variable which is set earlier in the process flow.
RESULT_CODE:
SCRIPT :
SUCESS_THRESHOLD : 0
by examining (RUNTIME CONTROL CENTER SCHEMA).all_rt_audit_execution_params
I can see that the variable contents assigned to RESULT_CODE look like this
'/usr/bin/rsh -n WINSRV1 dtsrun /usa /ppw /S WINSQLSRV /N JOBNAME_SHORT'
or this
'/usr/bin/rsh -n WINSRV1 dtsrun /usa /ppw /S WINSQLSRV /N JOBNAME_LONG'
I put the -n in because I thought that may have been the problem, but I think its more likely that there is come sort of ".properties" in a runtime file that needs to be set to control the "timeout" of user defined activities.
(like in Runtime.properties
property.RuntimePlatform.0.NativeExecution.Shell.security_constraint = NATIVE_JAVA
except maybe
.........Nativeexecution.timeout = .....?????
Has anybody had a problem like this? Found a document describing available configuration properties for PF activites?
Can anyone help?

I have been using OWB (10g r2) for a while and have used "User Defined" activities in production process flows successfully before.
However my problem now is that the activity (shell command) I want to perform, takes more than 10 minutes to complete. If I run the process flow with a parameter that generates a shell command that executes in two or three minutes, the process flow continues past the sucessfully executed user defined activity to the next activity.
BUT
If I run the same process flow (without modification or even redeployment) with a different parameter , the user defined acivity sucessfully executes (after about 11 minutes or so) but the process flow , doesnt seem to know that it is finished and just says it is still executing.
my "user defined" activity has
COMMAND: /usr/bin/ksh
PARAMETER_LIST bound to a variable which is set earlier in the process flow.
RESULT_CODE:
SCRIPT :
SUCESS_THRESHOLD : 0
by examining (RUNTIME CONTROL CENTER SCHEMA).all_rt_audit_execution_params
I can see that the variable contents assigned to RESULT_CODE look like this
'/usr/bin/rsh -n WINSRV1 dtsrun /usa /ppw /S WINSQLSRV /N JOBNAME_SHORT'
or this
'/usr/bin/rsh -n WINSRV1 dtsrun /usa /ppw /S WINSQLSRV /N JOBNAME_LONG'
I put the -n in because I thought that may have been the problem, but I think its more likely that there is come sort of ".properties" in a runtime file that needs to be set to control the "timeout" of user defined activities.
(like in Runtime.properties
property.RuntimePlatform.0.NativeExecution.Shell.security_constraint = NATIVE_JAVA
except maybe
.........Nativeexecution.timeout = .....?????
Has anybody had a problem like this? Found a document describing available configuration properties for PF activites?
Can anyone help?

Problem with execution of process flow

Hello! I'm new in OWB, and I hope you will help me.
I started execution of process flow and then connection to my server failed.
The problem is: In Runtime Audit Browser all mappings are complete, but process flow is in "Busy" status. I waited for 24h, and nothing changed. (execution of mappings costed 4h).
what should I do?

Hi,
Use any SQL tool like TOAD and see from the backend if the OWB process is running, if it is still running, kill it. Due to server connection failure the OWB process might have been killed but the Runtime Browser might show it as busy.
Go the Workflow Repository Schema and run the following query:
update wf_item_activity_statuses set activity_status = 'COMPLETE'
where item_type = <Your Process Flow Package Name> and actvity_status != 'COMPLETE'.
This should solve the problem.
Regards
-AP

User defined activity problem in process flow

Hi ,
any one can help me out
I have added a user defined activity in process flow , and given command values as \\windows\system32\cmd.exe
and parameter list : ${input.value}
and in script : kpi.bat
where i written
cd \kpi_sap
for /f "tokens=1-5 delims=/ " %%d in ("%date%") do copy "c:\kpi_sap\kpi_sap.csv" %%e-%%f-%%g.dat
but its not working ...
can any one give me direction to get success in this activity
best regards,
mahi

Hi Mahi,
I see that your post is a few days old and you might have resolved this issue, but if you haven't you might want to change the command value to c:\windows\system32\cmd.exe instead. Its been a while since I worked with commands, but I remember that I had a lot of problems when I defined paths with '\\'. What I found out was that OWB jobs was run under the windows system account and the system account don't have access to networkshares (even if its really not on the network).
I can't say whether the script in the kpi.bat is wrong, but as long as you can run it manually using cmd.exe you should be ok.
If this doesn't help it might be helpful with some more specific error messages.
regards Ragnar

Problems using user defined activitys in Process Flows

Hi everybody,
i have problems using an user defined activity in my owb process flow.
I did the following according to https://blogs.oracle.com/warehousebuilder/entry/how_to_use_user_defined_activity_in_owb_process_flow.
1. create a script in /tmp called test.sh
2. chmod +x on the script so it can be executed
3. Changed the Runtime.properties file under ORACLE_HOME/owb/bin/admin
property.RuntimePlatform.0.NativeExecution.FTP.security_constraint = NATIVE_JAVA
property.RuntimePlatform.0.NativeExecution.Shell.security_constraint = NATIVE_JAVA
property.RuntimePlatform.0.NativeExecution.SQLPlus.security_constraint = NATIVE_JAVA
property.RuntimePlatform.0.NativeExecution.OMBPlus.security_constraint = NATIVE_JAVA
4. Creating an Process Flow with just the user defined activity
5. Deploying the Process flow was successfull
But everytime i execute the pf i get the following error in my Job Details.
TEST_PF:USER_DEFINED
Parameters
Log file is available. Click Information button to view.
It says Log file is available. But i can't find this file. Also i don't know what the problem is.
Can anybody help me out of this?
Thanks a lot

Hi,
solved the problem by myself.
I had problems with my script. So i insert set -x into the script an executed the pf again. Know i can see whats happening in a log. You can find this log in the job detail window by pressing the blue ! Button when the whole PF (the first entry) is marked.
Greetings

Major problem - Process Flows hanging randomly - 11.2.0.1

Similar Messages

Maybe you are looking for