Client Copy FAQ
Which data has been actually copied with a client copy?
The data is selected through copying profiles.
Is it possible for me to only copy the Customizing and subsequently receive the application data in the target client?
No, this is not possible.
The existence of mixed number ranges, further adds to the problem and there is no solution. For eg, for both Customizing and application addresses are visible in the central address management.
It is possible to copy amid systems with unlike R/3 versions/add-ons?
The main or Basis version should be similar. In both systems, the support Packages and ADD-ONS also should have the same status. That’s when the consistency of the data be guaranteed.
For instance, the remote copy will terminate with a reference to differences in the dictionary or missing tables in the analysis phase if an add-on is only installed in the source system, and a copy will not be made. The remote copy dismisses incase all data has not been copied, which means a field is missing in the target system or if a field is too small or a complete table is absent. The Individual tables and, from version WAS 6.10, also development packages (formerly development classes) in the source system in an exception list, to avoid the termination, can also be included.
The target system is not checked during the export as export does not experience this problem with the transport. All the differences in the dictionary or missing tables are then reported as errors during the import however not resulting in any termination. All this data can be imported is imported. Therefore, the various ADD-ONs would not create a problem for the client transport. To what extent is the data then consistent or useable?
Can the client copies in a heterogeneous system landscape be created?
Remote copies or client transports can be made amongst various operating systems and database.
Between a client transport or a remote copy, which is faster and better?
Before version 4. 6, there are no major differences related to the runtime. The benefit of a client transport is that the source client is available again for users very soon after exporting the data, while with the remote copy, when the copy has been completed, you are only permitted to work in the source client.
Besides these, during the client transport you have a 'frozen' dataset of the client which cannot be altered, and which can be accessed or restarted repeatedly during the import.
Before version WAS 6. 10, a client transport is more error-tolerant where there are alterations in the table definitions (DDIC). Before WAS 6.10, only individual tables could be ignored in the source system incase this is included in the exception list in the source system (doc 70290). As of version WAS 6.10, in the exception list, a complete development packages (formerly development classes) can now be included. There is also a special option which avoids termination.
With parallel processes from version 4. 6, the remote copy is quick than in a client transport. Generally, the entire copying time for a remote copy is much below the pure export time of a client transport. However, you need a LAN link. You should check very critically whether the link does not slow down the copy too much, with a remote copy via a WAN.
Where can I find the cause of an error or a copy termination?
doc 22514 lists that data required for error analysis along with the location of this data. In addition, the syslog in both the target system and source system, which you can call by utilizing the transaction SM21, also delivers valuable information.
Why does a remote copy terminate, referring to DDIC differences?
The DDIC (data dictionary) belongs to the repository of the system (programs, system variables etc.). A repository is generally demanded on the same development level, for a client copy.
A remote copy dismisses incase not all data could be copied, eg when an entire able is missing or a field is missing in the target system, a field is too small.
Why do we have differences between the systems, for example, with the programs, even after I copy with the copy profile SAP_ALL?
The functional scope of the tools of the client copy has been explained in doc 24853. A client copy is only capable of copying customizing and possibly application data. Repository (for example, programs and table definitions) and other cross-client data is not copied and should reach the system via the corresponding transport routes.
What is the use of client 001 ?
During the initial installation of a system Client 001 is formed as a practice client or a reference client. Apart from this feature, there is no other difference. Hence, it is dealt with in the same way like every other customer client.
The Solution Manager up to Version 3. 2 is an exception. In this case, client 000 does not have Customizing for setting up the service desk. You are required to use client 001 as a copy template. The error is corrected with Solution Manager 4.0.
One more exception is NetWeaver as of Version 2004S. With a new installation, a client 001 is created for providing storage for the Java users. You can copy the users from client 001 with a client copy, if required. For Customizing, it is still suggested that you use client 000 as a copy template. This is applicable for especially to all products which run on NetWeaver (such as ECC). There are a few incorrect installation guides still available. If possible, you must obtain the latest version.
You may have to log onto the J2EE Engine and set the ume.r3.connection.master.client parameter on the new PI/XI client, after the copy
Can client 001 be deleted ?
Yes. (Except for the Solution Manager before Version 4.0. If required you should first copy the Java users.)
Can client 001 in production be used ?
For setting up a new client, it is suggested to customize a copy from client 000. Apart from this, a customer client should also be set up without any other application data, and the client 001 becomes seems more old with every Support Package and with every language installation.
However, incase you have already used client 001 for production, then there is no reason for not retaining this client.
Can more than one client for production in a system be used?
Yes. There are certain restrictions, however.
There seem to be issues, with number ranges in the target client after a client copy. Why does this happen?
During the copy, work was carried out in the source or target client. The client is varying and the copy should be repeated.