RADE

The Underlying Database Management System

The SCM system relies on a DBMS for storing data. Actually there are two DBMS that are supported by SCM: Oracle, DB2.

Whatever the DBMS is, the SCM installation falls in three steps:

1. Create an empty database
   There is no tool provided with SCM for creating a database. The database administrator has to use the DBMS's tools to do it.
2. Reference this database in SCM's profiles
   Register the name of the new database and specified the paths to DBMS client libraries.
3. Create a database schema
   Use SCM's SQL programs to create account and tables needed by SCM.

The reader has to go through all sections to achieve the installation.If there are some differences between Oracle or DB2, it will be mentioned.

[Top]

Oracle Server Installation

The Oracle installation is done by a Database Administrator (DBA) on the SCM server.
One database is enough for a small development project.

• Refer to the Oracle documentation for learning how to create a new database.
• Read this first section before creating this database.

One Oracle instance must be created. SCM will reference this instance by its SID
SCM provides SQL programs for

• Creating Oracle “table spaces”
• Creating an Oracle account for the SCM administrator
• Creating tables of the SCM database schema
• Running statistics on a database to get best performances.

Those programs will be generated by the adl_admin_site command (see "Customizing the Oracle Instance")

SCM does not provide any program for creating an Oracle instance. This must be done using Oracle tools.
Note however that specific “table spaces” will be created by SCM SQL programs upon the newly created Oracle instance.

SCM supports the "WE8ISO8859P1" character set. This character set is specify when creating a new database. It is recommended to verify that the new instance has been created with the same character set.

This characteristic can be checked on an existing database by running the query « select * from V$NLS_PARAMETERS » It should return WE8ISO8859P1 as value for 2 parameters.

PARAMETER                               VALUE                                                           
--------------------------------------- ------------
NLS_LANGUAGE                            AMERICAN                                                        
NLS_TERRITORY                           AMERICA                                                         
NLS_CURRENCY                            $                                                               
NLS_ISO_CURRENCY                        AMERICA                                                         
NLS_NUMERIC_CHARACTERS                  .,                                                              
NLS_CALENDAR                            GREGORIAN                                                       
NLS_DATE_FORMAT                         DD-MON-YY                                                       
NLS_DATE_LANGUAGE                       AMERICAN                                                        
NLS_CHARACTERSET                        WE8ISO8859P1
NLS_SORT                                BINARY                                                          
NLS_NCHAR_CHARACTERSET                  AL16UTF16

[Top]

DB2 Server Installation

The DB2 installation is done by a Database Administrator (DBA) on the SCM server.
One database is enough for a small development project.

• Refer to the DB2 documentation for learning how to create a new database.
• Read this first section before creating this database.

One DB2 database must be created. SCM will reference this database by its logical name
SCM provides SQL programs for

• Creating tables of the SCM database schema
• Running statistics on a database to get best performances.

Those programs will be generated by the adl_admin_site command (see "Customizing the Database")

SCM does not provide any program for creating a DB2 database. This must be done using DB2 tools.
Here are some advices to help to create a database suitable for SCM.

SQL Sample:
create database SCMDB(*) ... collate using identity

SQL Sample:
create database SCMDB(*) using codeset <codeset> territory <territory> ...

SQL Sample:
GRANT DBADM ON DATABASE TO USER adl(**)

db2set command:
db2set DB2_RR_TO_RS=YES

SQL sample:
update database configuration for SCMDB(*) using LOCKLIST 1000

SQL Sample:
update database configuration for SCMDB(*) using applheapsz 256 app_ctl_heap_sz 512

(*) The databasename SCMDB is given as an example. Replace it by your own database name.
(**) Refer to the DB2 documentation for the rights that can be given to a user.

[Top]

CD-ROM Installation on the SCM Server

In the rest of this document, INSTALL_PATH will refer to the path of the installation directory (given when running the installation program)
Once the installation is finished, tools are found under INSTALL_PATH/OS where OS is a directory the name of which is an acronym that depends on the operating system. Refer to [6] to get the current list of operating systems supported and associated acronyms.

To access SCM commands, you will run

On Unix

. INSTALL_PATH/xxx_a/code/command/tck_init

Note the ‘.’ at the beginning

On Windows

INSTALL_PATH\intel_a\code\command\tck_init.bat

then

tck_profile SCM_adm_V5R12_T12

or

tck_profile V5R12_T12

The first TCK is dedicated to the SCM administration and the second one is for all users.
Do not use them right now because a site profile must be filled in before; this is the purpose of the next section.

TCKs' suffixes that appear in this article might not be those you are using (they depend on the version of CAA-RADE CD-ROM). However the way to install SCM remains the same.

[Top]

SCM Initialization Profiles

There are two kinds of profiles: user and administrator profiles.
The administrator profiles gives access to all user commands plus specific administrative commands.
The administrator profile is used

• for launching the SCM Monitor
• for launching the SCM File Server
• for launching the Multi-site Transfer Manager
• for accessing administrative commands such as adl_mk_site, adl_mk_tree, adl_mk_root_ws, etc

The two profiles are generated when unloading the CD-ROM. However they call a “site” profile to set information related to the current installation.
The profile path is

• INSTALL_PATH/OS/code/command/admin/site_profile.sh for UNIX platforms
• INSTALL_PATH\OS\code\command\admin\site_profile.bat for Windows platform.

Where OS is a directory the name of which is an acronym that depends on the operating system. Refer to [6] to get the current list of operating systems supported and associated acronyms.

The definition of the user and password to connect to SCM databases will be generated by the adl_admin_site command (see "SCM Databases Connection").

This "site" profile must be edited according to the current installation.
Here is what is generated by default:

rem Following variables must be set up for current site
set ADL_TMP=C:\TEMP 
set ADL_TRACE_DIR=<TRACE_DIR> 
set ADL_CENTRAL_DB_ID= 
set ADL_MONITOR_DB_ID= 
set ADL_MONITOR_PORT= 
set ADL_MONITOR_HOSTNAME=
set ADL_DBMS=
set ADL_JAVA_ROOT_PATH=
set ADL_DOC=
rem Following variables are needed for the underlying DBMS
set ADL_LIB_DBMS=
set ADL_ORA_TNS_ADMIN=
set ADL_DB2INSTANCE=

# Following variables must be set up for current site 
export ADL_TMP=${ADL_TMP:-/tmp} 
export ADL_TRACE_DIR=${ADL_TRACE_DIR:-$HOME/.Adele/log} 
export ADL_CENTRAL_DB_ID=  
export ADL_MONITOR_DB_ID=  
export ADL_MONITOR_PORT= 
export ADL_MONITOR_HOSTNAME=
export ADL_DBMS=
export ADL_JAVA_ROOT_PATH=
export ADL_DOC=
# Following variables are needed for the underlying DBMS
export ADL_LIB_DBMS= 
export ADL_ORACLE_HOME=
export ADL_ORA_NLS33=  
export ADL_ORA_TNS_ADMIN=
export ADL_DB2INSTANCE=

ADL_TMP must be set with the path to the directory in which SCM will create temporary files.

ADL_TRACE_DIR must be set with the path to a directory in which all SCM command traces will be stored.
Each time the SCM profile will be executed, a trace cleaner will be launched and will remove all traces older than 7 days.

ADL_CENTRAL_DB_ID must be set with the SCM database name: it is the name that has been chosen when creating the database dedicated to the SCM product.

• it is the Oracle SID if you use Oracle as the DBMS.
• it is the logical name specified when creating the database if DB2 is your DBMS.

ADL_MONITOR_DB_ID must be set with the SCM database name.

ADL_MONITOR_PORT must be set with the port number for communicating with the SCM monitor. We advise to use a port greater than 1024 (lower numbers are reserved for system programs).

ADL_MONITOR_HOSTNAME must be set with the name of the machine where SCM daemons will be running (SCM  Monitor, File Server).

ADL_JAVA_ROOT_PATH: some SCM commands are graphical tools (adl_gui, graphical merger, etc) that require the "java" language to be installed on the user host. Use the ADL_JAVA_ROOT_PATH variable to specify the java location. For Instance, if the java program path is "C:\Program Files\IBM\Java13\jre\bin\java.exe" then the ADL_JAVA_ROOT_PATH variable must be set  with "C:\Program Files\IBM\Java13\jre".
Note that due to an Oracle bug, the path to the java program must not contain a closing brace ")". If it is the case (for instance on windows XP 64 bits machines where the 32 bits java runtime is installed into the following path : "c:\program files (x86)\java\jre...."), you must re-install the java runtime so that its path does not contain a closing brace.

ADL_DOC must be set with the path of the home page of the SCM documentation.

ADL_DBMS must be set with either "Oracle" or "DB2" according to your DBMS.

ADL_LIB_DBMS is a path indicating where the DBMS client libraries are located. It can be a single path or several paths separated with ';' on Windows or ':' on UNIX.

Here are some explanations for finding out what are the different values:

Once the “site” profile has been correctly updated, SCM commands are available after having executed either the user or the administrator profile. Note that they are not executed directly, these profiles are declared in appropriate TCKs (Tool Configuration Key).  The tck_init profile is used for setting up the current environment to give access to TCK commands.

On Windows, run:

INSTALL_PATH\OS\code\command\tck_init.bat

On UNIX, run:

. INSTALL_PATH/OS/code/command/tck_init

Where OS is a directory the name of which is an acronym that depends on the operating system. Refer to [6] to get the current list of operating systems supported and associated acronyms.

Then the tck_list command allows to display all available TCKs

tck_list SCM_adm_V5R12_T12 SCM_adm_V5R12_T12 V5R12_T12 Mkmk_V5R12_T12 SCM_V5R12_T12

Normally, after having unloaded the SCM CD-ROM, this tck_list command should display 2 TCKs (or 2 new TCKs) whose identifiers are:

• V5R12_T12: it corresponds to the tools found on the CD-ROM and will be the TCK to be used by all users
• SCM_adm_V5R12_T12: it is an administrative TCK and give access to SCM administrator commands (in addition to user commands)

Examples on one UNIX platform
Here are the commands any user must execute to access SCM commands:

. INSTALL_PATH/xxx_a/code/command/tck_init tck_profile V5R12_T12

Here are the commands to execute to access SCM administrator commands:

. INSTALL_PATH/xxx_a/code/command/tck_init tck_profile SCM_adm_V5R12_T12

Remark: as TCK identifiers are built from CATIA release (V5R12, V5R13, etc) and build version (T12, T13, etc), it is possible that the real identifiers are not those used in this document.

[Top]

Customizing the Database for SCM

SCM provides a program for generating SQL scripts that will be executed on the SCM database (see variables ADL_CENTRAL_DB_ID and ADL_MONITOR_DB_ID in the “site” profile that was edited in the previous installation step).

The database will contain three sets of tables that will be created by three different SQL scripts.

• Using Oracle, these tables will be created in « table spaces » that are defined by a fourth SQL script
• Using DB2, we do not provide any SQL script for creating "table spaces" since they have been created when creating the database.
  Using DB2, only one "table space" will be used for storing SCM tables. The name of this "table space" is the default proposed by DB2, that is "USERSPACE1". If you do not want to use this name, you will have to edit the SQL files and to replace each occurrence of "USERSPACE1" by your own "table space" name.

To access the SQL generator program, one must use the administrator TCK  SCM_adm_V5R12_T12 that gives access to administrator commands and especially to the adl_admin_site command. This command generates 5 kinds of SQL programs for Oracle and 4 kinds of SQL programs for DB2.

• InitInstanceXXX.sql is used for creating  « table spaces » and the Oracle account for the SCM administrator (Oracle only).
• CreateTablesXXX.sql is used for creating tables in the « table space(s) »
• DropTablesXXX.sql is used for deleting tables (only in case of restarting an installation)
• EstimateStatisticsXXX.sql must be run regularly on the SCM database in order to keep good performances
• Oracle.<date>.txt or Db2.<date>.txt is used by the adl_upgrade_db command for migrating an existing database.

[Top]

Using Oracle

the InitInstanceXXX.sql program requires parameters.

• First parameter = user to connect to database. Use "adl" if you don't defined your own user.
• Second parameter = password to connect to database. Use "adl" if you don't defined your own password.
• Third parameter = path to the directory where the program will create four « table spaces » used for storing persistent data (tables).
• Fourth parameter = path to the directory where the program will create four « table spaces » used for storing persistent data (indexes).
• Fifth parameter = path to the directory where the program will create a « table space » for storing persistent data.

Example:

mkdir /SCM_DB/TBS_DATA /SCM_DB/TBS_IDX /SCM_DB/TBS_TMP

InitInstanceXXX.sql db_user db_password /SCM_DB/TBS_DATA /SCM_DB/TBS_IDX /SCM_DB/TBS_TMP

For performance considerations it is advised to locate (at least) /SCM_DB/TBS_DATA and /SCM_DB/TBS_IDX on different disks.

Only one InitInstanceXXX.sql program must be executed on the Oracle instance – whatever it is (Monitor, cdb , dept).
Conversely the 3 CreateTablesXXX.sql programs must be executed.
InitInstanceXXX.sql must be used first, then execute the three CreateTablesXXX.sql programs.

Only the InitInstanceXXX.sql script requires parameters.

[Top]

Using DB2

The CreateTablesXXX.sql does not require any parameter. SCM tables are to be created in the DB2 default "table space" named "USERSPACE1". If you do not want to use this default table space, edit the SQL files and replace the "USERSPACE1" string by your own "table space" name.

Example:

CreateTableXXX.sql

The 3 CreateTablesXXX.sql programs must be executed even one datatabase is used for storing the three kinds of data (monitor, cdb, dept).

[Top]

How to Generate the SQL Programs?

We advise to use “small” as database size for the following commands if we consider that we are going to create a database for managing at most 50 workspaces and up to 5000 files in a workspace.

Run the three following commands to generate all the SQL programs (UNIX samples):

MONITOR

adl_admin_site –db SID –db_user adl –db_type monitor –db_size small –init_db –target_dir /tmp

CDB (Central Database)

adl_admin_site –db SID –db_user adl –db_type cdb –db_size small –init_db –target_dir /tmp

DEPT (Departmental Database)

adl_admin_site –db SID –db_user adl –db_type dept –db_size small –init_db –target_dir /tmp

How to Use the SQL Programs?

Connect to the target instance using its logical name and being the administrator and execute the SQL programs:

• when working with Oracle, the sqlplus program can be used for executing the SQL programs
• when working with DB2, the db2 program can be used for executing the SQL programs

Examples using sqlplus (found in the working environment)

mkdir /SCM_DB/TBS_DATA /SCM_DB/TBS_IDX /SCM_DB/TBS_TMP cd /tmp sqlplus system/manager@SID @InitInstanceMonitor.sql adl adl /SCM_DB/TBS_DATA /SCM_DB/TBS_IDX /SCM_DB/TBS_TMP ...

• system is the Oracle administrator login
• manager is the password for the “system” login
• SID is the SID of the Oracle instance to connect to
• CreateTablesMonitor.sql is the program to execute
• adl adl /SCM_DB/TBS_DATA /SCM_DB/TBS_IDX /SCM_DB/TBS_TMP are the parameters of the program to execute

sqlplus adl/adl@SID @CreateTablesMonitor.sql … sqlplus adl/adl@SID @CreateTablesCdb.sql … sqlplus adl/adl@SID @CreateTablesDept.sql

Note: the two last programs may produce errors indicating that some tables already exist. Do not worry about them because the 3 last programs are designed for being executed on 3 different databases and then some tables are duplicated between those databases. When only one database is used for SCM, there is no duplication and the same table cannot be created more than one time.

Examples using the db2 application (found in the working environment)

db2 connect to <database_name> user <login_name> using <password> ... db2 -tvf /tmp/CreateTablesMonitor.sql ... db2 -tvf /tmp/CreateTablesCdb.sql ... db2 -tvf /tmp/CreateTablesDept.sql

[Top]

Migrating an Existing SCM Database

This section explains how to migrate an existing SCM database in order to manage it with a new SCM version. You can skip this step if you are starting with SCM.

SCM provides the adl_upgrade_db command whose purpose is to connect to the existing databases and to learn about their current levels. Depending on these levels, the command will execute some SQL programs to migrate the database tables to a level suitable for the current version of SCM. A migration falls down in two steps:

• generate the SQL programs
• run adl_upgrade_db which will use the SQL programs

V5R8 specific note: if you are upgrading a database that has been created by a V5R7 version of SCM, it is needed to start the SCM File Server(s) before running the adl_upgrade_db command because it needs to communicate with it. 

Take care that you cannot go back from a migration except by restoring a backup of your database. Note also that the whole SCM site must be managed by the same version of SCM. For example, you cannot have a department database managed by a version of SCM while the central database is managed by another version of SCM (except if they are compatible, that is to say there is no migration to do between the two versions).

[Top]

How to Generate the SQL Programs?

Like for a database initialization they are generated by the adl_admin_site command. However less parameters are needed since the databases exist and are referenced in the current working environment.
The parameters are:

• the database administrator
• the size of the SCM project (can be small | medium | large)

adl_admin_site –db_user adl –db_size small –init_db –target_dir /tmp

Once the command is terminated, the adl_upgrade_db command can be launched by specifying the <-target_dir> directory in which it will look for SQL files. The command will connect to the central database and will retrieve the list of department databases to be upgraded too. The command will automatically compare the current level of each database with the new level and will apply the appropriate SQL programs to reach this level.

Once the migration will be started, the previous version of SCM won't be used anymore. It is advised to backup all databases before starting the migration.

adl_upgrade_db -dir /tmp ...

[Top]

Structure of Migration Files

The files generated by the adl_admin_site command can be edited but the following syntax must be respected:

• lines between [] cannot be changed
• lines starting with -- are comments
• other lines are SQL commands

$ more /tmp/Oracle.2001_10_08.txt [Central + Departmental Step 1] [Central Step 1] [Departmental Step 2] -- Add new columns in workspace table [Departmental Step 3] -- Drop columns in multi_tree_ws table RENAME adl.multi_tree_ws TO adl.multi_tree_ws1; [Departmental Step 3] CREATE TABLE adl.multi_tree_ws ( id VARCHAR2(20) NOT NULL, case_name VARCHAR2(32) NOT NULL, lower_name VARCHAR2(32) NOT NULL, is_deleted VARCHAR2(1) NOT NULL, nb_linked_ws NUMBER(10) NOT NULL, native_database VARCHAR2(20) NOT NULL, merge_cmd_end VARCHAR2(1) NOT NULL, case_tck VARCHAR2(32) , lower_tck VARCHAR2(32) , creation_hist_evt VARCHAR2(20) NOT NULL ) PCTFREE 10 PCTUSED 70 ...

SQL programs can be seen in order to let the database administrator knows what will be changed during the migration. Obviously any changes must be done with care to avoid any crash of SCM commands afterwards.

[Top]

Launching the SCM Monitor

The SCM monitor must always be running and is needed by all commands.
Here are the commands to execute for launching the SCM Monitor.

[Top]

UNIX Platform

1/ Set up the working environment for accessing administrator commands

. INSTALL_PATH/xxx_a/code/command/tck_init tck_profile SCM_adm_V5R12_T12

2/ Launch the SCM Monitor

adl_monitor –port $ADL_MONITOR_PORT –db $ADL_MONITOR_DB_ID

• $ADL_MONITOR_PORT is the port number specified in the “site” profile
• $ADL_MONITOR_DB_ID is the name of the database managing Monitor’s information and which has been specified in the “site” profile

3/ To check that the SCM Monitor is running, you can use the adl_mo_admin command:

adl_mo_admin -ping Pinging host ... in port: ... Replay from ... time 2ms ... #ERR# ADLMo - 0120: The connection with monitor on host: ... and port number: ... has been successfully tested.

[Top]

Windows Platform

Like on UNIX it is possible to launch the SCM Monitor by opening a terminal window (Command Prompt), setting up the environment and running the adl_monitor command. Here is an example of the commands to be used:

INSTALL_PATH\intel_a\code\command\tck_init.bat tck_profile SCM_adm_V5R12_T12 adl_monitor -port %ADL_MONITOR_PORT% -db %ADL_MONITOR_DB_ID% > C:\TEMP\Log_Monitor.txt 2>&1

However this way of running the Monitor needs to open a window and to let it open, otherwise the Monitor would be killed by the system.

To run the Monitor in background like a daemon on UNIX, it is possible to create a Windows Service for the Monitor. Consult the Administration article for learning about SCM services [2].

[Top]

Declaring the Site

Actually there is only one site that can be declared in SCM but it must be and this is the first information to be entered before starting with SCM. This is done using the adl_mk_site command and specifying a string like an identifier. Note that this site identifier will be displayed afterwards to identify the origin of object changes (in case of transfer of changes between different sites).

Example:

adl_mk_site DS_Headquarters

[Top]

Declaring the Database(s)

This step consists in declaring to SCM the name(s) of the database(s) that will be used for managing SCM data. You have at least one database to declare, the one specified in the environment variable ADL_CENTRAL_DB_ID. In the case where only one database is used for managing all SCM data, it must be declared as being as the same time a "central", a "monitor" and a "departmental" database ; this is done by the following command (UNIX sample)

adl_mk_db $ADL_CENTRAL_DB_ID -central -monitor -dept

If you are upgrading a database that have been used with a version of SCM older than V5R11, the adl_upgrade_db command should have declared automatically the existing database(s). In that case, use the adl_ls_db command to display the characteristics and use the adl_set_db command to change them if they don't correspond exactly to your installation.

[Top]

Setting Up and Launching the SCM File Server

Before to launch the SCM File Server, it is needed to

1. declare it into the SCM database
2. format the dedicated repository

[Top]

UNIX Platform

1/ Set up the working environment for accessing administrator commands

. INSTALL_PATH/xxx_a/code/command/tck_init tck_profile SCM_adm_V5R12_T12

2/ Declare the File Server

adl_mk_contents_server -cs_host <hostname> -cs_port <port_number> -desc <a short description for identifying this File Server>

• <hostname> is the name of the current host
• <port_number> must be a port greater than 1024 and not already used by another process (of course it must be different than the one chosen for the SCM Monitor).

3/ Format the File Server repository

The file server stores its data in two directories that must be local to the SCM server These directories must be created empty then the following command must be executed:

adl_file_server -data <path_to_the_first_directory> -files <path_to_the_second_directory> -format -port <port_number>

• -format option is used the first time to specify that the two directories are empty and must be formatted for further use.
• <port_number> is the port chosen when declaring the File Server.

Here is an example of the use of this command:

$> mkdir /SCM/FileServer $> mkdir /SCM/FileServer/data /SCM/FileServer/files $> adl_file_server -port 23451 –data /SCM/FileServer/data -files /SCM/FileServer/files -format $>

4/ Start the File Server

Once the environment has been set up, the way the File Server is started is the same as when formatting its repository except that the -format option is not used.

adl_file_server -data <path_to_the_first_directory> -files <path_to_the_second_directory> -port <port_number>

5/ To check that the File Server is ready to respond

Use the adl_fs_admin command with the -ping option:

adl_fs_admin -ping -port <port_number>

The command does not display any message and return 0 as error code when the file Server is running, otherwise a error message is displayed and the error code is greater than 0.

[Top]

Windows Platform

Like for the UNIX platform, the File Server must be registered using the adl_mk_contents_server command, then the repository directories must be formatted and finally the File Server can be launched using the adl_file_server command.

However it is advised to use a Windows service for launching the File Server since using a terminal window needs that this window remains open, otherwise the File Server would be killed by the system. Consult the "Administration" article for learning about SCM services [2].

Here are the steps to go through on Windows platform.

1/ Set up the working environment for accessing administrator commands

. INSTALL_PATH\intel__a\code\command\tck_init.bat tck_profile SCM_adm_V5R12_T12

2/ Declare the File Server

adl_mk_contents_server -cs_host <hostname> -cs_port <port_number> -desc <a short description for identifying this File Server>

• <hostname> is the name of the current host
• <port_number> must be a port greater than 1024 and not already used by another process (of course it must be different than the one chosen for the SCM Monitor).

3/ Format the File Server repository

The file server stores its data in two directories that must be local to the SCM server These directories must be created empty then the following command must be executed:

mkdir <path_to_the_first_directory> mkdir <path_to_the_second_directory> adl_file_server -data <path_to_the_first_directory> -files <path_to_the_second_directory> -format -port <port_number>

• -format option is used the first time to specify that the two directories are empty and must be formatted for further use.
• <port_number> is the port chosen when declaring the File Server.

Here is an example of the use of this command:

$> mkdir E:\SCM\FileServer $> mkdir E:\SCM\FileServer\data E:\SCM\FileServer\files $> adl_file_server -port 23451 –data E:\SCM\FileServer\data -files E:\SCM\FileServer\files -format $>

4/ Start the File Server

Once the environment has been set up, the way the File Server is started is the same as when formatting its repository except that the -format option is not used.

adl_file_server -data <path_to_the_first_directory> -files <path_to_the_second_directory> -port <port_number>

5/ To check that the File Server is ready to respond

Use the adl_fs_admin command with the -ping option:

adl_fs_admin -ping -port <port_number>

The command does not display any message and return 0 as error code when the file Server is running, otherwise a error message is displayed and the error code is greater than 0.

[Top]

Launching the Multi-site Transfer Manager

The Multi-Site Transfer Manager is needed only when transferring data between different SCM installations.
Here are the commands to execute for launching the Transfer Manager.

Running Account and Multi-site Server

Whatever the current operating system is, the account used for starting the SCM Multi-site Transfer Manager is the one that will be used when running SCM commands afterwards. For that reason, we advise the SCM administrator to use an account that is granted for managing SCM workspaces but which is not the SCM administrator account.

In addition, the manager accepts commands other than SCM commands (see adl_transfer_client [5]). To reduce the risk of executing undesirable commands, it is possible to install the manager on a specific host with limited access, for instance by refusing connections on any port but the one used for starting it. The default list of authorized commands is given in the file INSTALL_PATH/xxx_a/code/command/admin/AuthorizedCommands.txt.

[Top]

UNIX Platform

1/ Set up the working environment for accessing administrator commands

. INSTALL_PATH/xxx_a/code/command/tck_init tck_profile SCM_adm_V5R12_T12

2/ Use the adl_start_transfer_mngr.sh shell for starting the Multi-site Transfer Manager (the port used here is given as an example)

adl_start_transfer_mngr.sh –port 23452

[Top]

Windows Platform

Like on UNIX it is possible to launch the Multi-site Transfer Manager by opening a terminal window (Command Prompt), setting up the environment and running the adl_transfer_mngr command. Here is an example of the commands to be used:

INSTALL_PATH\intel_a\code\command\tck_init.bat tck_profile SCM_adm_V5R12_T12 adl_transfer_mngr -port 23452 -cmdfile %ADL_ADMIN_PATH%\AuthorizedCommands.txt

However this way of running the manager needs to open a window and to let it open, otherwise the process would be killed by the system.

To run the Multi-site Transfer Manager in background like a daemon on UNIX, it is possible to create for it a Windows Service. Consult the Administration article for learning about SCM services [2].

[Top]

Creating the First Workspace Tree

Once the Monitor and the File Server started, we can create the first tree using the adl_mk_tree command.

Once the tree created, we must create the first workspace that will be the root workspace of the workspace tree. We must use the adl_mk_root_ws for creating the first workspace (the others will be creating using the adl_mk_ws command).

The adl_mk_tree and adl_mk_root_ws are considered as administrative commands and thus are available only when executing the command tck_profile SCM_adm_V5R12_T12, i.e. the TCK giving access to administrator commands

Some help is available for each command by running the command with the –h option, for instance:

adl_mk_tree –h ...

Refer to the “Workspace Management” paper [3] in the CAA-RADE V5 online documentation for learning more about creating and populating a workspace.

[Top]

Using the EstimateStatisticsXXX.sql Programs

These files have been generated using the adl_admin_site command when customizing the database instance for SCM.

• EstimateStatisticsMonitor.sql for the database that stores information for the monitor
• EstimateStatisticsCdb.sql for the database that stores site information
• EstimateStatisticsDept.sql for the department database

The database administrator must execute these files once a week in order to

• keep good performances of the database (for both Oracle and DB2)
• prevent a wrong result from some queries when using Oracle 8.0.5

  [Top]

Installing a SCM Client Host

Installing the SCM product on a client consists in downloading the CD-ROM, then editing the "site" profile like on the SCM server.

In fact the main thing to take care of is the location of the DBMS client libraries. If they are located in the same directories as for the SCM server, the server's "site" profile can be reused for clients, otherwise another one must be defined for specifying the new locations.

Each user must execute tck_init, then tck_profile to access SCM user commands.

UNIX Platform

. INSTALL_PATH/OS/code/command/tck_init tck_profile V5R12_T12

Where OS is a directory the name of which is an acronym that depends on the operating system. Refer to [6] to get the current list of operating systems supported and associated acronyms.

Windows Platform

INSTALL_PATH\OS\code\command\tck_init.bat tck_profile V5R12_T12

Where OS is a directory the name of which is an acronym that depends on the operating system. Refer to [6] to get the current list of operating systems supported and associated acronyms.

[Top]

Troubleshooting

If either the monitor, the file server or the transfer manager seems to have some problems, traces can be found in the temporary directory of the host where they have been started.

On UNIX platform

All daemons print traces directly in the window where they are started. It's up to the SCM administrator to decide to redirect them into a dedicated trace file.

On Windows platform

Where the daemons are started using services, traces are automatically redirected into specific files that are located in the directory specified by the variable ADL_TMP. Under this directory, there is a specific trace file for each daemon:

• for the monitor, the trace file is "Log_SCM_Monitor_<TCK>.txt"  where <TCK> is the corresponding TCK (for instance V5R12_T12)
• for the file server, the trace file is "Log_SCM_FCS_<TCK>.txt" where <TCK> is the corresponding TCK (for instance V5R12_T12)
• for the transfer manager, the trace file is "Log_SCM_TMANAGER_<TCK>.txt" where <TCK> is the corresponding TCK (for instance V5R12_T12)

For example, to see the traces of the SCM monitor on Windows platform:

E:\> INSTALL_PATH\intel_a\code\command\tck_init.bat E:\> tck_profile SCM_adm_V5R12_T12 E:\> cd /d %ADL_TMP% C:\TEMP\> type Log_SCM_TMANAGER_V5R12_T12.txt ...

[Top]

In Short

The SCM product is installed on the SCM server machine in four main steps:

1. customizing the SCM profile by editing the "site" profile
2. customizing the database instance by executing SQL programs
3. launching the Monitor and File Server daemons
4. creating the first tree and root workspace

Once the server side installation completed, the client side can be installed on each client host by downloading the CD-ROM and editing the "site" profile according to the database used and the characteristics of the SC installation on the server side.

[Top]

References

[Top]

Copyright © 2000, Dassault Systèmes. All rights reserved.