Nos encontramos en algunos casos que una cambio de Service pack o una actualización grave sobre producción representa una incompatibilidad entre el trabajo de los usuarios y las tareas de desarrollo, sin llegar a exagerar como las instalaciones que hacíamos cuando heramos esclavos de las consultoras y a la “grupy” financiera de turno se le ocurría de debíamos tener el entorno de producción, el de preproducción y el de desarrollo. No es mala idea tener un lugar en el que poder trabajar con libertad tanto para desarrollo como para pruebas funcionales.

Axapta básicamente esta compuesto por un código fuente que se almacena en unos directorios dentro de unos archivos específicos y una base de datos. Todo esto no funciona sin un programa que interprete estos archivos y los enlace con la base de datos.

De esta manera para hacer un entorno de desarrollo a partir de producción solo debemos mantener esta coherencia directorio base de datos.

Ejemplo practico.

Tenemos una instalación tipo, 3.0 spX con una base de datos SQL server X.

Nuestra instalación se encuentra situada en “C:\Navision\Axapta Application\Appl” dentro del directorio tenemos los directorios de las diferentes aplicaciones. Normalmente Standard y producción.

Copiando producción y renombrar el directorio como Test ya tendríamos la mitad de nuestro entorno de desarrollo.

Después solo hay que realizar una copia de seguridad de la base de datos completa y restaurarla con otro nombre.

El último paso seria configurar el cliente o el AOS para que accedan al nuevo entorno y sincronicen los posibles cambios.

Introducimos el jscript que nos proporciona el servidor de analisis y actualizamos la version.

When setting up an Oracle database to work with Axapta it is often (normally) configured as a dedicated database server for the following reasons:

· Until now this has been the default behavior of an Oracle database.

· Early versions of the multithreaded server was known to be unstable and with some annoying bugs.

The backside of configuring the database as a dedicated server is:

· Each new database connection creates a process / thread on the server.

· Each new database connection allocates more memory for PGA purposes than a similar multithreaded connection.

As you increase the number of concurrent users in an Axapta system you will also increase the administrative burden of the Oracle database server, because you will start new processes / threads on the database server for each new Axapta session. At one point this task of switching between processes / threads will be so overwhelming big part of the servers activity that it will “kill performance” on the server.

Large Oracle Axapta installation with several hundred concurrent Axapta session are therefore more or less forced to configure their database as a multi threaded server.

Axapta running in 2-tier mode has in average 3-4 database connection, which is why configuring your Oracle database as multithreaded will ease the database servers tasks of administer the processes / threads and decrease the consumption of memory. Even though Axapta in 2-tier mode is more expensive database connection wise compared to 3-tier mode you will still benefit from using multithreaded server when running Axapta in 3-tier mode.

This document describes in short how to your Oracle database as a multithreaded server.

1. Oracle init.ora parameters

If you want to understand the way a multithreaded server works compared to a dedicated server, then you can read about this in the following 2 Oracle manuals:

· Oracle Concepts

· Oracle Administrator guide

The following Oracle kernel parameters must or can be specified in order to make the Oracle database server run in multithreaded server mode. A detailed description of these parameters can be found in the “Oracle Reference” manual.

Parameters  
mts_dispatchers:   Controls the protocol and number of dispatcher processes started at instance start. In the most simple way you just specify the protocol and the number of dispatcher processes stated at instance start:
·    mts_dispatcher=”(protocol=<protocol>)(dispatchers=<n>)”

<protocol> would typically be TCP and <n> is the number of dispatchers. You can specify multiple MTS_DISPATCHERS in the initialization file, but they must be adjacent to each other.

To specify that you want to start 5 dispatchers for the TCP/IP protocol at instance start you must specify as follows:
·    mts_dispatcher=”(protocol=TCP)(dispatchers=5)”

mts_max_dispatchers:  Specifies the maximum number of dispatcher processes allowed to be running simultaneously. Default: 5 or the number of dispatchers configured, whichever is greater. For most systems, a value of 200 – 250 connections per dispatcher provides good performance.
mts_servers :  Specifies the number of server processes that you want to create when an instance is started up. If system load decreases, this minimum number of servers is maintained. Therefore, you should take care not to set mts_servers too high at system startup.
mts_max_servers:   Specifies the maximum number of shared server processes allowed to be running simultaneously. If artificial deadlocks occur too frequently on your system, you should increase the value.
Default: Derived from mts_servers (either 20 or 2*mts_servers) .
Tests in out benchmark lap shows that in order to avoid artificial deadlocks in a 3-tier environment you should specify mts_max_servers as 1.2 * Number of active Axapta sessions. If you have 500 concurrent Axapta sessions, then you should specify mts_max_servers to be 600.
instance_name :  Name of the instance. Default value is the instance’s SID. NB! Due to a bug in Oracle 8.1.7 you must specify this parameter in order to run multithreaded server.

2. Start of multithreaded database environment

After you has changed the Oracle initialization parameters in your init.ora file, then you must bounce the database (shutdown and start again). This you have to do to let the dispatcher processes be registered by the listener process and thereby start the database as multithreaded. When running a multithreaded database you must always start the listener processes before you start the database. If the database is started before the listener, then the database will start the dispatcher processes, but because the listener don’t know their existence the database will run as a dedicated server.

Check in your alert log for your database that the dispatchers has started. Just after the start banners of the database and non-default system parameters you should find 2 lines looking something like this:

starting up 5 shared server(s) …

starting up 2 dispatcher(s) for network address ‘(ADDRESS=(PARTIAL=YES)(PROTOCOL=tcp))’…

To make sure that the listener has registered the dispatcher processes correctly start the listener control utility and list services supported by the listener (bold text indicates the commands you have to execute):

C:\>lsnrctl

LSNRCTL for 32-bit Windows: Version 8.1.7.0.0 – Production on 14-MAY-2001 12:37:27

(c) Copyright 1998 Oracle Corporation. All rights reserved.

Welcome to LSNRCTL, type “help” for information.

LSNRCTL> set password oracle

The command completed successfully

LSNRCTL> services

Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=wk0-kma-dev-nt4)(PORT=1

21)))

Services Summary…

AXDB has 3 service handler(s)

DEDICATED SERVER established:0 refused:0

LOCAL SERVER

DISPATCHER established:0 refused:0 current:0 max:1022 state:ready

D001 <machine: WK0-KMA-DEV-NT4, pid: 338>

(ADDRESS=(PROTOCOL=tcp)(HOST=wk0-kma-dev-nt4.intern.dd.dk)(PORT=1196))

DISPATCHER established:0 refused:0 current:0 max:1022 state:ready

D000 <machine: WK0-KMA-DEV-NT4, pid: 337>

(ADDRESS=(PROTOCOL=tcp)(HOST=wk0-kma-dev-nt4.intern.dd.dk)(PORT=1195))

The command completed successfully

In the case above is the listener named “listener” (the default value) and there is no database explicit specified in the listener SID_LIST_<listener> in the listener.ora file.

To check that the connection is established in shared mode you can query services again and check if the established counter for one of the dispatchers has been increased.
The listener.ora file used in the above shown example look like:

# LISTENER.ORA Network Configuration File: D:\ORACLE\NETWORK\ADMIN\listener.ora

# Generated by Oracle configuration tools.

LISTENER =

(DESCRIPTION_LIST =

(DESCRIPTION =

(ADDRESS = (PROTOCOL = TCP)(HOST = wk0-kma-dev-nt4)(PORT = 1521))

)

(DESCRIPTION =

(ADDRESS = (PROTOCOL = IPC)(KEY = WK0KMA))

)

)

PASSWORDS_LISTENER= (oracle)

From Axapta point of view you don’t have to change anything to utilize the new setup at the database unless you for some reason wants to connect to the database in dedicated mode. To do so you have make a new entry in your tnsnames.ora file where you (SERVER=DEDICATED) in your CONNECT_DATA section for that entry.

AXDBTCP.WORLD = (DESCRIPTION =

(ADDRESS_LIST =

(ADDRESS = (PROTOCOL = TCP)(Host = dbserver)(Port = 1521))

)

(CONNECT_DATA = (SID = AXDB)(SERVER=DEDICATED))

)

Be aware that for Oracle databases running on a Windows server is TCP/IP the only supported protocol for multithreaded connections.

axl_gdi_2.xpo

Nada de soluciones Salomónicas de update a “cholones” contra la custinvoicejour y cosas así que se os ve la pinta de “programatas” de medio pelo salidos de la vieja escuela del “clipper 5″, ahora el famoso “Delfi 6″. (jejejejej)

Bueno lo dicho aquí va el código para contabilizar facturas o albaranes con la numeración que queramos sin que el estándar de deje de chutar.

protected void insertJournal()
{;
ttsbegin;

numberSeq = this.allocateNumAndVoucher();
//axl numero de factura..
if (salesParmUpdate.invoiceid != ”)
{
number = salesParmUpdate.invoiceid;
numberSeq.parmNumberSequenceCode(”);
ttsbegin;
voucher = numberSeq.voucher();
ttscommit;
}
else
{
[number, voucher] = this.getNumAndVoucher();
}

if (this.updateNow())
{
this.postUpdate();

TransactionLog::create(this.transactionLogType(),this.transactionLogTxt());
ttscommit;
}
else
throw error(strFmt(“@SYS21533″));
}

Tan solo hay que sobre escribir el método insertjournal de la salesformletter

static server void Start()
{
asciiio file;
str txt;
;
winapi::deleteFile(‘C:\\file2.txt’);

file = new asciiio(‘c:\\filex.bat’,'w’);

if (file)
{
txt = (strfmt(‘net start “Axapta batch 1″ ‘));
file.write(txt);

file = null;

winapi::shellExecute(‘c:\\filex.bat’);
winapi::deleteFile(‘C:\\filex.txt’);
}

}

ftx_tpv_shutdownrestartservices.xpo

Axapta system tables contain data that the Axapta kernel uses to control the behaviour of the kernel.  Axapta system tables are created automatically by the kernel, when Axapta is started for the first time against an empty database or after a service pack upgrade.

Axapta system tables are in many ways treated as ordinary Axapta tables from the kernel’s point of view. They are synchronized in same way and inserts / updates of the tables are done in a similar way.

System tables are not company specific and do not have a “DataAreaId” column.  Therefore, system tables are implicitly treated as if they belong to the “DAT” company.

All system tables (except “SqlSystemVariables”) have a “RecId” column, which is updated as an ordinary table belonging to the “DAT” company.

Some system tables has the following standard columns:

·         MODIFIEDDATE

·         MODIFIEDTIME  

·         MODIFIEDBY    

·         CREATEDDATE   

·         CREATEDTIME   

·         CREATEDBY

All system tables (except “SqlSystemVariables”) have a TableId.  The TableId’s for system tables starts at 65508 and each field in a system table also has a Id that follows the normal rules for Id generation of FieldId’s.  Please refer to the form “SysSqlDictionary” to see the mapping of system tables and TableId / FieldId.

Axapta system tables can’t be accessed or viewed directly from the AOT (Axapta Object Tree).  Furthermore, you should not access the system tables Axapta application layer.

This document contains a description of the most important Axapta system tables, except the table “SqlSystemVariables” which is described in another document. The information in this document is not valid for the table “SqlSystemVariables”.

Description of Axapta System Tables

ACCESSRIGHTSLIST (TableId = 65526)

This table contains info about explicit granted access rights for the defined user groups.

Note: Until you have explicitly defined permissions for a table, access to tables is defined using Feature keys. Once you set restrictions on a particular table for a particular user group, all other groups automatically have No Access to that table. As this will seriously impact the way the system works, you are presented with a warning dialogue before any change takes effect. In the dialogue you must decide whether other user groups should have no access as the result of the change you are making, or if other groups should keep their present permissions.

COMMON

COMPANYDOMAINLIST (TableId =  65509)

This table contains relation between companies and domains.

DATAAREA (TableId = 65533)

This table contains data about created companies.

DATABASELOG (TableId = 65508)

This table contains data about which tables and events to log. The events are logged in the Axapta table SysDatabaseLog.

DOMAININFO (TableId =  65510)

This table contains data about domains.

PRINTJOBHEADER (TableId = 65525)

This table is used in connection with printing from Axapta.

PRINTJOBPAGES (TableId = 65524)

This table is used in connection with printing from Axapta.

SQLDESCRIBE (TableId = 65527)

This table is used temporarily during check / synchronisation process and should normally be empty.

SQLDICTIONARY (TableId = 65518)

This table contains data dictionary data about Axapta application and system tables (except SqlSystemVariables) as stored in Axapta data dictionary. Content of this table can be viewed by the form “SysSqlDictionary”

SQLPARAMETERS (TableId = 65517)

This table is for future use.

SQLSTATISTICS (TableId = 65513)

This table is for future use. It is planned to store Axapta specific statistic data, which can help the kernel to optimize the parse, bind, execute and fetch operations.

SQLSTORAGE (TableId = 65515)

This table contains Oracle storage parameters as specified in Axapta “Storage Setup” utility. The value are added to the CREATE statement, whenever a database object is created from Axapta.

SQLSYSTEMVARIABLES

The SQLSystemVariables table is not viewable from with Axapta; however, you can look at the contents by looking at the Database Information form.
SYSCONFIG (TableId = 65514)

This table contains license information, version, feature key setup and other basic system configuration data.

SYSLASTVALUE (TableId = 65528)

This table is used to store set values for specific program elements for individual users, which should survive from session to session.  This table is used by both the kernel and the application.

SYSTEMRECORDLEVELSECURITY (New for version 3.0)

This table contains info about record level security based on table, user and company. The stored information is the extra where condition for the table.

SYSSEMAPHORE

This is a virtual system table that is created and maintained by the kernel. The table can’t be found on the database; however, it contains data about all kernel semaphores

SYSTEMSEQUENCES (TableId = 65516)

This table contains sequence numbers for RecId and TransactionId. A row for each used company and sequence.

TABLECOLLECTIONLIST (TableId = 65532)

This table contains data about relation between virtual companies and table collections

USERGROUPINFO (TableId = 65530)

Table contains name for the defined user groups

USERGROUPLIST (TableId = 65529)

Table describes relation between users and user groups.

USERINFO ((TableId = 65531)

Table is used to store general set up values for individual users. Use the “Tool/Options..” menu-item  to change values in this table. If “UserInfo” is empty or entry for current user doesn’t exists, then no menu item or toolbar is available from Axapta. 

UTIL*  (UTILAPPLHELP, UTILCODEDOC,  UTILELEMENTS,  UTILELEMENTSOLD,  UTILIDELEMENTS, UTILIDELEMENTSOLD, UTILLOCKS)

These are all virtual tables containing info about programming elements.

VIRTUALDATAAREALIST (TableId = 65534)

This table contains data about relation between real companies and virtual companies.

Total  Ax (axapta 3.0) permite instalar un balanceador de conexiones entre dos o mas AOS, para que sirve esto, pues más bien para no mucho, el concepto es que el dependiendo del número de usuario que se han conectado a un u otro AOS este cluster distribuye la conexión entrante para mantener un numero homogéneo de conexiones. 

Nada de  transacciones o de carga de procesos, simple y “molt macu” para vender un poco de tecno-logia al puro estilo de las modas de internet. Seguro que en un par de meses se les ocurre algo para dual core.. jajjajaa..

Para terminar un copy paste de como se instala y se configura.. Perdon por los derechos de autor.. jejjeje

Axapta Object Server cluster

The Axapta Object Server cluster environment is build to provide load balancing and fall back. An Axapta Object Server can be specified to belong to a specific cluster. If the same cluster name is given for a number of object servers they all belong to the same cluster. A user can be specified to connect to a given cluster.

Suppose a number of servers are set up in the same cluster – named AOSC. They should all use the same application and database, and in other aspects have the same settings. All clients connecting to the AOSC should be specified to search for servers belonging to this cluster, otherwise the load balancing mechanism will not work properly. A client starting up against an Axapta Object Server Cluster will start out by broadcasting an UDP package requesting Axapta Object Servers fulfilling names and IP server masks given in the client settings to respond. It is therefore vital that firewalls allows such packages to pass through. The client will receive answers from all Object Servers with information on object server name and the name of the instance plus the number of clients connected to that instance. The client disregard information from servers not belonging to the requested cluster. The client then chooses one of the instances with the lowest number of clients.

Since a client automatically can choose from all servers in a cluster one or more Object Servers can be down with out changing the connection procedures. However performance might suffer severely due to higher load on the fewer servers.

Setting up the server

On the server manager select the server of interest and click settings. In the advanced field write -cluster=<my cluster name>. The cluster name can be up to 10 characters long.

Setting up the client

In the configuration tool choose the configuration of interest. In the advanced field write -cluster=<my cluster name>. The client setting for connecting to an object server depends on two situations. The first situation being both client and Axapta Object Servers being in the same subnet. The second being client and Axapta Object Servers belong to different subnets.

Recall the settings in the configuration utility for connecting to an object server:

- Axapta Object Server Mask
- Axapta Object Server host names
- Axapta Object Server IP Address masks

When the client and Object Servers belong to the same subnet the above three masks can be left blank.

When the client and Object Servers belong to different subnets the IP Address masks should contain the mask for the subnet the Object Servers belong to. An example could be 192.88.253.255, where the last 255 specifies that the search is done among servers in the subnet 192.88.253.*.

Note – if client at any time receives only service response from Object Servers within the same cluster it will automatically choose the least loaded of these servers without having to specify the cluster name in the client configuration. This means that if your installation is one single Axapta cluster, there is no need to change the existing configuration with respect to cluster name. Ensure that all relevant Object Servers are located by client configurations network set up.

Los formularios, informes y consultas utilizan  la clase Args como primer argumento en el constructor.

Pasandole el nombre del objeto a crear a una clase args y despues llamar a classfactory con estos argumentos podemos inicializar fomularios, informes etc.

    Args a = new Args(“CustTable”);
    formRun fr = ClassFactory.FormRunClass(a);
    fr.init();
    fr.run();
 

para recuperar el objeto args en la clase de destino, por regla general existe un metodo que nos devuelve este dato.

elment.args() en el caso de formulario.

El objetivo de esto es poder pasar a traves de args objetos completos a otros contextos de ejecución de forma que podriamos abrir un formulario inicial F1 ejecutar otro formulario secundario F2 y poder ejecutar metodos del formulario 1 desde el formulario 2.

un ejemplo.

Hemos creado un from con un metodo para recuperar el texto de un strignedit y este texto lo vamos a pintar desde un report, utilizando para ello una clase intermedia o el metodo del form que lo ha llamado.

fir_dem.xpo 

public void init()
{
    object ob;
    FIR_ReportClas FIR_ReportClas;

    super();

    //axl si no es la clase entonces es el formmm…
    if (classIdGet(element.args().caller()) == classNum(FIR_ReportClas))
    {
            FIR_ReportClas = element.args().caller();
            print FIR_ReportClas.texto();
    }
    else
    {
        ob = element.args().caller();
        print ob.texto();
    }

os_authent_prefix  

Con un prefijo tipo ‘ops$’ nos sevira,  de manera que oracle sepa identificar si el usuario del SO que solicita la conexión esta creado en oracle autorizando el acceso.  Crearemos un usuario  en oracle como este:

-- Windows

CREATE USER "OPS$DOMINIO.COM\DBO" IDENTIFIED EXTERNALLY;

GRANT CONNECT TO "OPS$DOMINIO.COM\DBO";

Crearemos un usuario en nuestro dominio en este caso DBO .Oracle –> OPS$dominio\DBOWin –> DBOY podremos finalizar la instalación.Ni que decir tiene que debemos crear un tablespace de datos otro de indices etc. (Vamos lo típico en estos casos).http://www.oracle-base.com/articles/misc/OsAuthentication.php Para finalizar recomendaros que ORACLE + SUSE Linux  + AX es la mejor combinación para un optimo rendimiento de la instalación y evitar los sustos que nos dará Microsoft con el día a día que ya sabemos todos.  Un saludo.