Setting up the Empress Distributed Server consists of three steps:
For example, to set up a Distributed Server with the following configuration:
The Global Data Dictionary, which must be set up on the client (local) machine, is an ASCII file and is created by running the Empress utility program empsvutl with the syntax:
empsvutl -gdd GDD
where:
| GDD | is the name of the Global Data Dictionary file which can be any filename. |
Empress responds with the following:
Logical Database name <hit return to quit> :logicaldb1 Full path name of the Database in the Server machine (e.g. /usr/empress/db_name) > /usr/database/db1 Distributed Server name [LOCAL] : srv1 Logical Database name : logicaldb1 Physical Database path name : /usr/database/db1 Distributed Server name : srv1 Is this correct (y/n/q) ? y Logical Database name <hit return to quit> :
In this example the Global Data Dictionary is named GDD and resides in /usr/dbadmin.
The contents of the file are as follows:
MSDBLOGICALNAME=logicaldb1 MSDBLOCATION=/usr/databases/db1 MSDBSERVERNAME=srv1 MSDBLOGICALEND
Set the Empress environment variable MSGLOBALDATADICTIONARY to indicate the location of the Global Data Dictionary.
For example, in the UNIX C shell:
setenv MSGLOBALDATADICTIONARY /usr/dbadmin/GDD
or in the UNIX Bourne or Korn shell:
MSGLOBALDATADICTIONARY=/usr/dbadmin/GDD export MSGLOBALDATADICTIONARY
The Server Configuration File is an ASCII file and is created by running the Empress utility program empsvutl with the syntax:
empsvutl -scf SCF
| SCF | is the name of the Server Configuration File which can be any file name. |
Empress responds with the following:
Distributed Server name <hit return to quit> : srv1
Node name of the machine where Server resides: > copper
Number of Server sub-processes [1] :
Name of Server administrator [support] : joe
Note that : No two active Servers can have the same combination
of host name an Internet port number
Internet port number (5000 - 65534) : 5001
Distributed Server name : srv1
Server host name : copper
Number of Server sub-processes : 1
Name of Server administrator : joe
Network access type : INET_SOCKET
Internet port number : 5001
Is this correct (y/n/q) ? y
Distributed Server name <hit return to quit> :
For the purpose of this example, we will use the name SCF and place the file in the directory /usr/dbadmin. The contents of the file are as follows:
MSSERVERNAME=srv1
MSSERVERHOSTNAME=copper
MSSERVERNPROCS=1
MSSERVERADMINISTRATOR=joe
MSSERVERNETTYPE=INET_SOCKET
MSINETPORT=5001
MSSERVEREND
Set the Empress environment variable MSSERVERCONFIGFILE to indicate the location of the Server Configuration File on the server machine.
In the UNIX C shell:
setenv MSSERVERCONFIGFILE /usr/dbadmin/SCF
or in the UNIX Bourne or Korn shell:
MSSERVERCONFIGFILE=/usr/dbadmin/SCF export MSSERVERCONFIGFILE
Follow the same procedure to set up the Server Configuration File and the environment variable on the client machine. Ideally, copy the file from the server machine to the client. This will guarantee that the files are identical.
The Empress utility empsv is used from the server machine to start the Distributed Server. The syntax is as follows:
empsv server_name
where:
| server_name | is the name of the server which is defined in the Server Configuration File. |
In our example, we start the server by typing the command:
empsv srv1
The server must reside on the same machine as the database it will service. The server must be started from the machine where it will run. The variable MSSERVERCONFIGFILE must be set before this command is run so that the server can find its configuration file when it starts.
There should be one process for each server sub-process as specified by MSSERVERNPROCS within the Server Configuration File. There will be one extra process for the server spooler-scheduler.