Static and Dynamic Server Selection

Static Spool Server Selection

With static server selection, the device definition in the spool system specifies the SAP spool server that is to process output requests for the device.

This static server selection is used in the following cases:

· You have not defined any special server definitions for the spool servers in transaction SPAD ( ® Spool Server ® Spool server ® Create/Create Using Template).

· You have defined and/or classified the server in SPAD, but have not filled out the fields Mapping and Alternative Server.

In this way, you can use the classification and identification attributes of spool servers without compromising the printing of requests in the sequence in which they are generated.

Dynamic Spool Server Selection

By default, the selection of a spool server is static, that is, defined by the spool server specification in the device definition.

You can select a Spool Server dynamically using the server definition in transaction SPAD ( ® Spool Servers ® Spool server ® Create/Create Using Template) using load balancing, the alternative server, or both. In this case, if the spool server fails or is overloaded, the system searches the hierarchy of servers that are specified in the Mapping and Alternative Server. The server with the lowest output workload is then used for printing.

Dynamic server selection can include load balancing or simply the use of logical or alternate servers. If you are allowing dynamic server selection, then the selection of the spool server to process an output request is not solely determined by the device definition. For the server selection to be performed dynamically, the server must be defined in transaciton SPAD ( ® Spool Servers ® Spool server ® Create/Create Using Template) and fulfill one of the following conditions:

· Load balancing is activated for the server specified in the device definition

· An alternative server is specified for the server in the device definition.

· A logical server is specified for the spool server in the device definition.

Process of Dynamic Server Selection

While with static server selection, the spool server specified in the device definition is simply used, with dynamic server selection, the system must find the spool server with the lowest output workload. The following figure shows how the system does this:

As in static server selection, the spool system starts by looking in the definition of the target output device to find the primary spool server. If you have created a definition for this in transaction SPAD ( ® Spool Server ® Spool server ® Create/Create Using Template), the spool system uses the algorithm described below to determine the spool server to be used.

It is important to note that this server-selection algorithm is recursive. As the spool system traverses the server hierarchies defined in the logical server definition (transaction SPAD ® Spool Server ® Spool server ® Create/Create Using Template ® Mapping field) and in the alternate server definition (in the Alt. server field), it applies these rules at each new level of the hierarchy.

...

1. First check: Is there any relevant information in the spool server definition? If the spool server is not a logical server and does not have an alternative server, the server specified in the device definition is used. In this case, the server selection is static.

If this result occurs while searching a spool server hierarchy, then this server is returned (if it is running) by the search as a potential server to use for the output request. This is a terminal node in such a hierarchy.

2. Is load balancing activated for this server?

If yes, then the spool system searches both the logical server and alternate server definitions, as available, to find the spool server that has the shortest queue of spool requests waiting.

If both a logical server definition and an alternate server have been specified, then the spool system compares the best server found in each hierarchy. The server that has the least output load is returned by the search, regardless of which hierarchy it was found in.

3. If load balancing is not in effect, then the spool system checks whether the server is actually a logical server.

If yes, then the spool system searches the logical server hierarchy (in the Mapping field) for a spool server to use. If the spool system finds an active spool server, then this server is used. If none is found, then the spool system searches the alternate server hierarchy, if available. If an active spool server is found there, then it is used.

The spool system uses the first active spool server that it finds. There is no attempt to distribute the workload among available servers.

4. If the server is not a logical server, then the spool system checks that the server is running and has a spool work process. If these conditions are met, then the server is used for processing the output request.

If these conditions are not met, then the spool system checks to see if an alternate server has been defined for the server.

If there is an alternate server definition, then the spool system searches the alternate server hierarchy for a spool server to use.

5. If all of these possibilities for finding a spool server fail, then the spool system reports an error with the output request. No spool server could be found to process the request.

For a detailed example of server selection, see the Example: Spool Server Selection section.

Restrictions of Dynamic Server Selection with Local Printers

You can use dynamic server selection without restrictions only for output devices that are defined with remote access methods S and U or with the local OMS access method E. The reason for this is that the SAP System accesses remote printers using a network link. This network link is likely to be usable without adjustment by any spool server. Output requests for a remote printer can therefore be moved from one spool server to another without causing any problems.

Output Management System (OMS) printers are accessed by a print command that can be executed locally from any spool server. Requests for OMS printers can therefore also be moved from spool server to spool server without problems.

For local printers (access methods C and L), on the other hand, dynamic server selection requires some preparation at the operating system level. In a local connection, the SAP spool server passes an output request directly to the spooler or print manager in the host on which it is running.

Such an output request can be processed at another server only if the local spooler or print manager at the second spool server can identify the target output device. This means that the target output device must be defined under the same name in both local spoolers (in the host of the first spool server, and also in the second host). For all practical purposes, an output device with a local access method must be defined under the same name in all hosts on which an SAP spool server runs. Otherwise, switching an output request for such a device to another spool could generate the error Device unknown from the local spooler or print manager.

The graphic below gives an example.

In the example, the SAP System is to access printer P330, through a local connection, at a Microsoft Windows NT system called NT1. The Print Manager of NT1 knows P330 under the name NT330. NT330 is entered as the host printer in the P330 device definition.

The local connection ­‑ using access method C ‑ means that the primary SAP R/3 spool server for printer P330 runs on NT1. The spool server is called NT1_C11_01.

However, server NT1_C11_01 is overloaded. NT1_C11_02 on NT2 is specified as the alternative server. Output requests for P330 can however be processed by the spool server on NT2 only if NT2’s print manager has a printer definition with the identical name NT330. This NT330 printer definition must also connect to the same printer as NT330 in NT1.

You can use dynamic server selection with local printers only if:

· You can define each affected local printer under an identical name in each host system on which an SAP spool server runs. The spool server can find the printer name NT330 on every host

· You ensure that local printers have names that are unique across all of your SAP host systems LPT1 on one host identifies the same printer as LPT1 on another host

· You can use problem-free dynamic server selection for OMS devices (SAP R/3 devices defined with access method E) as long as the commands used by the ROMS and LOMSs can be executed from each of your spool servers.