Replication process halts and aborts

Hello all,
We have a an instance of NS Directory 4.15 running on solaris 2.6.
We have 1 Supplier and 15 Consumers.
During replication of 4 consumers, the replication process halts and aborts,
and error messages similar to the following messages appear:
[29/Oct/2001:14:32:02 +0200] - Automatic online replica
creation is turned off. Replication cannot continue until you manually
reinitialize the consumer from the replication admin screen. Aborting.
[29/Oct/2001:14:32:02 +0200] - consumer XXX.XXX.XXX.XXX/o=XXX.XXX:
REPLICATION HALTED
the above error messages occur.
Why is it happen?
What should I do to prevent it?
Please let me know.
Thanks,

Ehhh.. Real-time makes it significantly more complex. What versions of Oracle & Sybase do you have?
To achieve this level of "real-time" replication, you are realistically going to have to set up a message-passing system to send the data changes from Sybase to Oracle. If you had a window of at least a few minutes, you could use something like Sybase Replication Server or Oracle Heterogeneous Services to do the replication in batches.
At a high level, whenever a data change is made in Sybase, a message will need to be created and propagated to the Oracle database where the change is applied. You can roll this sort of thing yourself using JMS (Java Messaging System) since that is pretty standard-- you can configure Sybase to generate JMS messages and configure Oracle Advanced Queues to accept those messages. You would then need to write a process in Oracle that dequeued and processed the messages.
Sybase Real-Time Data Services may be helpful here. What isn't clear to me, though, is how much code you really end up not having to write vs. how much effort is required on the setup & maintenance side. Whether you go for a custom point-to-point solution or the message bus approach probably depends on whether this is a one-off project or whether this is something that your organization is going to want to leverage for multiple projects.
Justin
Distributed Database Consulting, Inc.
http://www.ddbcinc.com/askDDBC

Similar Messages

How to externally trigger the execution of TestStand with a start and abort button through a digital interface?

Hello,
I'm Currently evaluating TestStand as an alternative for an in-house developed Test Sequencer.
To start our own Test Sequencer we use a small box, connected to a DIO board. The box has a start button and an abort button. The box also has a pass, a fail and a running led.
The interface with this box is made via some digital lines of a PCI-DIO-96.
In our own Test Sequencer we have groups named Init, Run, Abort, Exit.
The Init group is executed at startup (only once). It is used to initialize all the HW and SW. -> I guess this is the Pre UUT Loop callback in TestStand.
The Exit group is only executed once at the end of the day when the application is terminated. It is used to free all the used hardware and cleanup all the resources. This probably is the Post UUT Loop Callback in TS.
When a product needs to be tested, the operator presses the "start" button which triggers our own Test Sequencer and the run and abort group are being executed (first the run group, afterwards the abort group).
When the product is being tested the "running led" of the little box lights up to indicate to the operator that the application is running. (only when the run and abort group are running)
The Run group has all the functional tests in it. (MainSequence)
The abort group is used to put everything back in it's original state after the test on this single product is done. (Post UUT)
When executing the tests and something goes wrong (operator gets stuck in a clamper, ...) the operator can still press the abort button and then the execution immediately jumps from the currently executing step in the run group to the first step of the abort group. So, when something goes wrong, immediately the abort group is called.
At the end of the run and abort group, if no errors occured, the "pass led" lights up. If one or more steps went wrong the "fail" led lights up.
This setup can also be used to test multiple product in parallel. At that time all the different parallel testers have such a small box which contains a "start" and "abort" button and a pass, fail and running led. (it is possible that they are all connected to the same PCI-DIO-96 board. )
My question:
Is it possible to do something similar like this in Teststand? If yes, is there an example available that shows me how to do this in TestStand? (externally trigger the execution of TestStand)
Typically, in the Init group (Pre UUT Loop) the digital interface box gets initialised.
In the Close group (Post UUT Loop) the digital interface box is taken out of scope.
Note: The PCI-DIO-96 board to which the digital interface box is connected will also be used in the rest of the developed application (MainSequence, ...)
What's really important for me is that I can create a process model that all the application developers in our organisation can use. I really don't want any of the application developers (limited software experience) to mess around in all the features of TestStand.
For them it's important that they just add the function DigitalInterfaceBoxInit () in the Pre UUT Loop Callback and the DigitalInterfaceBoxClose () in the Post UUT Loop Callback, and then everything works!
It is important that the appliation developers do not have to create any global variables, other functions, synchronisations, parallel sequences, ... in TestStand. All this needs to be done with a simple call to the DigitalInterfaceBoxInit function.
Thanks in advance for all the help!
Best Regards,
Dennis Vanotterdijk
[email protected]

Dennis-
Your application sounds very well suited to TestStand's abilities. I am also quite impressed with your knowledge of how TestStand's architecture is arranged when you are still just considering whether or not to use it.
I think that TestStand would work really well for you in this application. Like you mentioned it will provide you with a form of standardization for your application developers to work from. It also provides the flexibility for you to add your custom routines in many different places. TestStand also makes parallel and batch testing much easier in TS 2.0 so that you could develop one test for your product and execute different instances of it in parallel to test multiple products at once.
As for your specific question about how to c
ontrol TestStand using a DIO board. I think this is very feasible and should not be too difficult. Since TestStand provides you the ability to create a custom operator interface your operator interface could monitor the status of your DIO board and launch/abort executions based on the read values. Usually the executions are launched/aborted when a button on the GUI is pressed however, I do not see any thing different about basing the action on a DI signal vs a mouse click. I am sure your application is more involved than this high level description but from the sounds of it I think it is very possible to do with TestStand.
Based on my experience of building test systems with TestStand and your description of the application, I would feel very confident in using TestStand to achieve all the goals you mentioned. If you have further detailed questions on how one of your features might be implemented, feel free to contact one of our Application Engineers or email us at www.ni.com/ask
and we would be glad to help you.
Best regards,
Richard McDonell
National Instruments

Userexit/BAdi in BP replication process in R/3?

Hi All,
Is there any userexit/Badi is available in R/3 for BP replication process for CRM to R/3? It would be much appreciated if you provide the same.
Thanks,
Narsaiah

Hi All,
Thanks for your response, after my investigation I found that there is no issue in CRM o/b BDocs, issue is only when populating the BP in R/3. To populate this it's using BDC call transaction there it's not populating the Name1 and Name2 fields of BP(Customer) in XD01. And its using BAPI PI_BP_PROXY_BAPI_CUSTOMER in R/3 side to process i/b data. I'm not sure why its not populating in call transaction. It's using DEBITOR_BDCDATA to get BDC data with this FM I'm bit confused, as its not picking any screen fields for Name1, Name2 and Address. However its updating address not Name1 and Name2. Could you please someone advise on this? It would be greatly appreciated.
Thanks in advance.
Narsaiah

Replication of Materials and services from R/3 to SRM

Hello,
I would like to get some advice about replicating materials and services from R/3 46C to SRM 5.0.
I replicated the objects succesfully in Development and the Test environment but now i have to do it in a Productive environment.
We use about 40.000 material and 5000 services in R/3.
Im concerned about the timing of the process.
How long will this aproximately take to replicate this amount of data?
What will be the best time to start replicating?
Is there any inconvenience in replicating during the normal working hours?
What happens if a user is currently using a material during the replication process? Is the material replicated anyway?
Thanks for any advice.
Aart

Hi,
How long will this aproximately take to replicate this amount of data?
This depends on the Production System Resources and the processes available. But for the number you have projected here should not take more than 2 - 3 hours at the max.
What will be the best time to start replicating?
Is there any inconvenience in replicating during the normal working hours?
What happens if a user is currently using a material during the replication process? Is the material replicated anyway?
It better to perform the activity during the night / production outtage time. Ideally users should be allowed to use them after the replication process. So, please do not perform this during business hours or when there are end users in the system. You can just trigger the replciation in the night and it should be completed in few hours. Besides, if there are any background activities the workprocesses will be occupied (very common in business hours) and hence there will not be enough processes for the replication job. This will increase the time for replication by many folds.
Hope this solves your problem.
Regards
Kathirvel

Query regarding replication process in oracle 8i

While replicating the master database I need some data. For that I stopped the replication process but in that time the application using the master database was not able to modify the database content because stopping of replication process put the master definition in read only mode.
So my query is Does oracle 8i puts the master definition site in read only mode if someone stops the replication ? If yes then is there any way to avoid this?
Thanks

Hi
First of all why you are using so much old version (oracle 8i).
when we have synchronous oracle database replication (multi master replication) then before saving our transaction at source site it will communication at all the oracle database server by using database link if at one site it find any link breakage then it will not save transaction at the source site and that transaction will ber rollback. ( Mean in synchronous replication your database link will be up any all time. This method is used where there are heavy network resource or they have replication on LAN.
In Relication 90% people use Asynchronous replication
In asynchronous replication if database link become down then our transaction become pending in the queue when link become resoted then transaction are pushed one by one in the fifo (First in First our).to their destination site to all other master or master definition sites.
Second
When we stopp replication there are two methods to stop the replication
Method A is quised the replication group at master definition site. When you quised replication group then your all tables present in that replication becomes read only you can not perform DML statment on that tables at all sites. and it is only perform by DBA to perform some administration activities like include new table in a group or regererate replication support of existing tables.or add new replication site.
Method B we totally sotpped replication that now our this server will not be in replication further. for this we use store procedure. like this
connect repadmin/repadmin@hr
SQL> Execute DBMS_REPUTIL.REPLICATION_OFF;
PL/SQL procedure successfully completed.
If you wants to again continue replication at that server. after connection wtih replication administrator apply the following command
SQL> Execute DBMS_REPUTIL.REPLICATION_ON;
PL/SQL procedure successfully completed.
For further problem you can contact me any time
Best Regard
Abdul Hameed Malik
[email protected]
Islambad Pakistan
Edited by: Abdul Hameed on Feb 9, 2009 12:04 AM
Edited by: Abdul Hameed on Feb 9, 2009 12:05 AM

How to supervise Replication Process

Hello,
On Directory Server 5.2. How can i supervise the replication process ? I want to supervise automaticly the process : if it is running or not. The best way (if it's possible), without scan access log.
Can you help me

I have not used the Perl tools I mentioned below.
Someone had recommended the use of these two tools to check if the replication has done its job of maintaining LDAP data integrity, i.e. same content before master and slave, I think this is more complete than just checking if replication is "running".
http://www.openldap.org/lists/openldap-software/200501/msg00167.html
http://search.cpan.org/~gbarr/perl-ldap-0.3202/contrib/ldifsort.pl
http://search.cpan.org/~gbarr/perl-ldap-0.3202/contrib/ldifdiff.pl
Gary

Script for checking the replication process in Netscape Directory Sever 4.1

Hi,
We are using Netscape Directory Server 4.1.6 in our production environment. Where we want to know is there any script available to check the replication process. Or can we write a script to check the process.
Waiting for your replies at the earliest.
Thanks,
Syed A.

Hi Syed,
Replication state is very easy to see in 4.x and earlier. The last changenumber in the supplier changelog is visible (on the root DSE IIRC), and each consumer suffix has an operational attribute "copiedfrom" that tells you the name of the supplier, the database generation id, and the last change replayed from the supplier.
So a script that compares the results of
ldapsearch -h <supplier> -s base -b "" "objectclass=*" lastchangenumber
and
ldapsearch -h <consumer> -s base -b <base_suffix> "objectclass=*" copiedFrom
will show you if a replica has fallen out of sync. It won't tell you more than how many changes behind the replica is, though, because pre-5.x changenumbers are sequential integers, not timestamps.
However, the change that corresponds to the changenumber is clearly visible under the "cn=changelog" suffix. So if you look at the change itself you can see the timestamp on it.
For more info on the status of replication, you should be able to query the agreements on the supplier. Those live under the "Netscape Machine Data Suffix", which you can also find on the supplier's root DSE:
ldapsearch -h <supplier> -s base -b "" "objectclass=*" netscapemdsuffix
And of course the error logs will usually tell you if replication is failing outright.
I may be off on some of the attribute names, but if you do some investigating along these lines, you should be able to figure it out.

Detect finishing of replication process

Hello,
i intend to use a HANA instance as HANA-based accelerator to accelerate some time-costly db reading transactions, using a secondary db connect.
I'm examining the SAP LT Replication Server for real-time data replication to HANA, for having my whole process of read and update data untouched, but transform the costly selection statements to use a secondary db connection.
While playing around, i noticed that the replication process isn't handled in real-time, and subsequent reading transactions operate on the non-replicated data set.
In my use case, this is a crucial problem.
Is there an appliance to monitor the replication process, so I can detect the finishing of the process? Is there a signal that flags the finishing of the replication process?
Have you any other approach to disrupt and continue my transaction sequence, to assure to operate on the most actual data?
Thanks, Stephan

Hi Tobias,
thanks for your reply.
To me, real-time would imply, that, at the end of my db transaction, the applied data record is already replicated. Yet the LT replication server offers a real-time launched replication, but that's ok if i can deal with it in a smart way.
In my usage scenario, accelerate extreme transaction processing, i have two subsequent transactions, both operating on the same data record. Transaction one updates the data record on the primary application db, which is to be replicated to a HANA db. Transaction two selects the data record via secondary db connection from the HANA db and returns the pre-replicated data record. I have average latency times below 1 second on this scenario.
For me, it would be sufficient to pause my process until the replication process is finished, the overall acceleration due to the HANA select makes it worth, but i need some kind of signal to continue my process and perform the HANA select on the most recent data set.
Regards, Stephan

Archiver Issue| Automatic replication process | Archiver thread blocking

Hi,
Recently, we are facing a strange issue with Archiver auto replication process. Sometime, archiver thread is blocking and auto replication is stalling. We observed this issue after we apply latest OUCM patches downloded from Oracle support site.
As an work around, we are doing UCM restart.
Anybody got this kind off issue? We are following this up with Oracle in parallel. Any help regarding this is highly appreciated.
Gowtham J

What version/patch set/core update of UCM are you using?
Jonathan
http://redstonecontentsolutions.com
http://corecontentonly.com

Start the replication process manually 9i oid

There is problem replication of db entries in oid, i found replication process is not running on that db...
How to start replication process manually ? i tried this but did'n work?
porter-db1$oidctl server=oidrepld connect=ldaphou1 instance=1 flags="-p 389 -h dolly-db1" start
porter-db1$ps -ef | grep oidrepld
porter-db1$ps -ef | grep oid
oracle9i 1301 1 3 Aug 18 ? 3:46 oidmon connect=ldaphou1 start
oracle9i 1315 1301 0 Aug 18 ? 57:15 oidldapd connect=ldaphou1 -p 389 -debug 0 -i 1 -conf 0 -sport 636 -sslenable 2
oracle9i 1321 1 0 Aug 18 ? 1219:56 oidldapd

-> Check oidrepld.log
-> Also check $ORACLE_HOME/ldap/log/remtool.log
-> Make sure the 'configset' and 'instances are right when you are attempting to start replication.
-- Pramod Aravind

Diff between Error,Warning and Abort message

Hi Friends
Can any body tell the Diff between Error,Warning and Abort message
Thanks

Error - Its the abnormal behaviour which was not expected. Its means the process has been hualted at this point. Unless and untill you resolve this, you can not procedd further.
Warnings: Its just the effect of a process it may have on other processes. You can continue doing irrespective of the warnings.
Abort : to terminate a computer processing or data transfer activity.
Thanks..
Shambhu

Process Chain and CPS

Hello all,
We are currently having a problem with submitting process chains. When the chain runs it does not display any of steps in the process chain and where they are running from the job we submit (RSI_START_BW_CHAIN) So what we see is that the PC has been submited and completes in say 17seconds...however everything is still running on the system...for example our CIF PC runs for 3hrs but alls we are able to see is that the job executed from CPS with no problem. I've been digging around trying to find clear documenation about PC and CPS with little luck. Am I approaching the job submission incorrectly? We are on 7.0.3.

Hello,
You should be using the RSI_RUN_BW_CHAIN job to start process chains. Maybe you can try that first. For the rest things should be straight forward. Depending on the BW backend systems you might encounter some issues, with BW 7 the synchronization has changed and you would be better of using one of the latest 7.0.4 versions (SP6 has just been released).
Regards Gerben

Process Chain and e-mail configuration

We have configured a process chain e-mail message recipient by keying in the full recipient e-mail address in the message screen. This works successfully.
However, we would prefer to use the SAP User ID in the process chain message configuration. (The sap User Id would contain the external e-mail id)
Why: Because we do not want to have to reconfigure the process chain every time a user or an e-mail changes. We would like to use a generic user in the process chain configuration, and then only change the generic user e-mail when changes occur.
When trying to use the sap user id in the process chain no mail message is received.
We would like to seek input from folks in this forum who may have worked with a similar scenario and have developed a solution, or could provide suggestions on how to make this work.
Thanks in Advance

Barbara,
          Mine all work fine on WLS or Jserve, maybe you should include a code snippet??
          Barbara Singer wrote:
          > Hello:
          >
          > We have some simple servlets that need to process forms and bundle this
          > information up in a SMTP message. This is very basic stuff but we can not
          > seem to get this to work with WebLogic AppServer 4.51. I've built the
          > sample servlet /examples/servlets/MailServlet and this does not work either.
          > It also throws now exception. I can see it connect to the SMTP server ans
          > send the information but nothing comes through.
          >
          > I also read in your documentation that /utils/sendMail will be a deprecated
          > method in a future release. Is it still supported in 4.51?
          >
          > Any input, sample examples working, etc. is welcome.
          >
          > /bas
          Russell Castagnaro
          Chief Mentor
          SyncTank Solutions
          http://www.synctank.com
          Earth is the cradle of mankind; one does not remain in the cradle forever
          -Tsiolkovsky

Process chain and Control M

Hi Gurus
We are designing process chain for our BW solution.
We have identified the dependancies of various loads which includes flat file loads and loads from R/3.
Now we would like to control ERP and BW jobs by Control M .
I have the following questions:
1. Which portion of BW process chain do we need to plug in to Control M. ? Do we need to connect the Meta chain to control M or different process chains to Control M?
2.When we create Process chain for BW, what is the best practice? Is it good to create small chains and connect them via Meta chain? or we have to create small chains only?
3. Can we include R/3 extraction Jobs and V3 Jobs in our BW process chain? so that they will trigger when we run the process chain?
4. If we decide to run V 3 jobs after every half an hour on R3 side then how often we have to extract data to BW? and how to catch correct delta without missing single record?
I would appreciate if you can help me with your knowledge.
Thanks
Kris

Hi Kris,
Control M is a third party scheduling tool provided by BMC software.
1) You have option to either include a meta chain or individual chains in Control M .
2) It depends on how your data is extracted. If you have any dependecy jobs on sourcre system, then it is better to create small process chians and include them along with your dependecy jobs in Control M
3)you can include R/3 dependency jobs and V3 jobs in Control M. Once those jobs are finished you can run BW process chains via Control M.
4)You have several scheduling options availalbe in control M to take care of your BW and R/3 jobs.
V3 jobs collect data from R/3 application tables and fill R/3 delta queue(RSA7). when you run BW process chain it will collect all the data available from R/3 delta queue.
If both jobs are running at same time, then it is better to apply some wait time on BW jobs(in process chians).
hope this helps

Process Chain and Info Package Transportation?

Hello All
Case 1
How to transport a Process chain and a Info Package Individually?(first time from DEV to QA)
Case 2
I have 9(IP1 to IP9)infopackages and 3(P1 to P3) Process chains and in each Process chain I have 3 Info packages like IP1,IP2,IP3 in P1 and so on.,
Now is it necessary to transport Infopackages and Process chains individually or is it enough to transport Process chains only(If v transport Process chains will the Infopackages will also transport)then how to assign a request to each one?
What I mean is,all Infopackages are in $TMP(Package),and Process chain is also $TMP,Now is it necessary to change the package for Infopackages as well as Process chain and Tranport only PC,Iam confused here,can anyone help me out?
Many thanks
balaji

Hi Balaji,
If you have created new infopackages then you have to transport them. It is not necssary to transport the info-packages if you have them in the process chain but have not changed them.
Also you cannot transport any BW object with $ tmp package. Assign a package and then transport the objects.
The best way to transport the process chain is from the RSPC screen. Click the transport button there and collect all the objects and transport them.
Bye
Dinesh

Replication process halts and aborts

Similar Messages

Maybe you are looking for