Академический Документы
Профессиональный Документы
Культура Документы
Guide
BusinessManager 6.5
Clustering
General. This Agreement represents the complete agreement concerning the License granted hereunder and may be amended only, in whole or in part, by an addendum in writing executed by duly authorized representatives of Savvion. If any provision of this Agreement is held to be unenforceable, such provision shall be removed only to the extent necessary to make it enforceable. This Agreement shall be governed by the laws of the State of California.
Third Party Software. SBM provides a modified PNGencoder from http://www.keypoint.com/. Original com.keypoint.PngEncoder is a Java class that converts Image objects to PNG byte arrays. Copyright (C) 1999 J. David Eisenberg. This library is free. You can redistribute it and/or modify it under the terms of the GNU Library General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. This library is distributed for your use, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Library General Public License for more details. You should receive a copy of the GNU Library General Public License along with this library. If not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. SBM includes software developed by the jfor project (http://www.jfor.org). Copyright (c) 2002 by the jfor project. All rights reserved. For more information, go to http://www.jfor.org/jfor-license.html. SBM includes a JDBC driver from Oracle for Oracle database 9i/10g. (c) Copyright 2004, Oracle Corporation. All rights reserved. Review the license terms and agreements for the JDBC driver at http://otn.oracle.com/software/tech/java/sqlj_jdbc/htdocs/jdbc9201.html. Oracle is a trademark of the Oracle Corporation. SBM includes a jConnect 5.5 driver from Sybase. Review the license terms and agreements for the driver at http://www.sybase.com. (c) Copyright 2004 Sybase Inc. SBM provides an embedded application server, evaluation version, from Pramati. Copyright 2004 Pramati Technologies. All rights reserved. Review the copyright for Pramati app server at http://www.pramati.com/. Contact Savvion, Inc. to obtain full licenses for this application server. The source code, object code, and documentation in the com.oreilly.servlet package are copyrighted and owned by Jason Hunter. Binary code of some servlets in this package is included in SBM. Copyright (c) 1998-2004 Jason Hunter. All rights reserved. Review the copyright and agreements at http://www.servlets.com/copyright.html. SBM includes software developed by the Apache Software Foundation (http://www.apache.org/) and therefore is subject to the terms of the Apache Software License, Version 1.1 (http://www.apache.org/LICENSE). SBM includes Mozilla Rhino JavaScript implementation in Java. See Mozilla Public License at http://www.mozilla.org/MPL/MPL-1.1.html. Copyright (c) 1999-2006 by Savvion Incorporated, 5104 Old Ironsides Drive, Suite 205, Santa Clara, California 95054, USA.
Table of Contents
Preface
Audience....................................................................................................................................................... 7 Information on SBM Documentation.............................................................................................................8 Conventions Used in this Manual ................................................................................................................. 9 Product Support Contact Information ........................................................................................................... 9
Chapter 1Introduction
Clustering Terminology ...............................................................................................................................10 Designing Server Clusters.......................................................................................................................... 11
Web Cluster............................................................................................................................................. 24 Standalone Servers ................................................................................................................................. 26 Proxy Server............................................................................................................................................ 27 Document Server .................................................................................................................................... 28 URL for the Document Server .......................................................................................................... 28 Viewing the XML File...............................................................................................................................29 Running modifyprop Utility ...................................................................................................................... 29 Updating the JNDI Properties File...........................................................................................................29 Running the Cluster.................................................................................................................................... 29 Adding New and Recovered Nodes ...........................................................................................................30 Editing and Deleting Nodes ........................................................................................................................ 30
Configuring for QuickRules Support........................................................................................................ 69 Setting Up the Cluster ................................................................................................................................ 69 Configuring Portal Load Balancing.......................................................................................................... 69 Modifying SBM Properties .......................................................................................................................... 70 Configuring BizPulse for Failover ............................................................................................................... 71 BizPulse Application Loading in a Cluster............................................................................................... 71 Working with the Cluster............................................................................................................................. 71 Starting EJB and Portal Clusters ............................................................................................................. 71 Starting IBM HTTP Server ................................................................................................................ 72 Starting the BizLogic Server .................................................................................................................... 72 Starting the BizPulse Server.................................................................................................................... 72 Stopping the Cluster ................................................................................................................................ 73 Restoring Original Configuration ................................................................................................................ 73 Working with Two-Node Clusters ............................................................................................................... 74
Glossary Index
Clustering Guide
Preface
This guide is part of the documentation set for Savvion BusinessManager, version 6.5 SP2.
Audience
The range of Savvion BusinessManager (SBM) modules creates the opportunity for four types of users, summarized in the following table:
Table 1 SBM User Types and Responsibilities User Type SBM User Responsibilities Responsible for carrying out specific business tasks using SBM applications. Examples include: SBM Manager An office administrator submitting an online purchase request. HR personnel scheduling interviews for a potential hire. A sales representative responding to a customer request. An engineer reviewing a trouble ticket reported by a customer.
Responsible for automating business processes in a particular business domain. BPM Portals Management module serves as the primary interface to SBM for the SBM Manager, enabling the manager to monitor, analyze, and control business processes. Also uses the Process Modeler for modeling and simulation. Responsible for creating customized applications for implementing business processes and developing interfaces associated with tasks. SBM application developers may work closely with SBM business managers to define the requirements of an application, and determine the business processes. Responsible for configuring SBM components, managing user/group profiles, and administering SBM applications and utilities.
Application Developer
SBM Administrator
monospace
backward slash \ SBM_HOME or %SBM_HOME%
Your name, installation site address and the license key for Savvion software. Your SBM version and build number. Your operating system, application server and browser, with version and service pack details, if any. Your database management system and version, and information on JVM and JDBC used.
Table 4 Savvion BusinessManager Product Support Team Contact Information To contact by... Web Site Use... http://www.savvion.com/support/ If you are a first time user, you can send an e-mail to Product Support and request them to provide you with a user name and password to access this site. If you are already a customer, you can also access SBM discussion groups at: http://www.savvion.com/support/tsgsupport/discus/ E-mail Telephone support@savvion.com 1-877-SAVVION in the United States or 1-408-330-3000 for local calls 408-330-3111 for international calls 1-408-330-3444 for the Product Support Fax Line Postal Address Savvion BusinessManager Product Support Team 5104 Old Ironsides Drive, Suite 205 Santa Clara, CA 95054 USA
Chapter 1
Introduction
Clustering involves grouping independent nodes together to work as a single system that allows shared processing. A cluster is a loosely coupled group of nodes collaborating to provide shared access to the services that each node hosts. To the other systems, or to the other components of the same system, the cluster appears as a single system. The objectives of clustering are:
Providing high availability by balancing resource requests. Ensuring reliability through reduced system failure, by managing failover. Offering scalability by allowing dynamic addition of resources to improve performance.
Together, a clustered system provides higher availability, increased reliability, and greater scalability than each node can provide independently.
Clustering Terminology
Some important clustering terminology is explained below. Node. Refers to the individual system that is part of a cluster. Each node in the cluster will have its own local resources. However, the cluster also has common resources, such as a common data storage. These common resources are accessible to each node in the cluster. Load Balancing. Is the ability to switch between the locations where similar requests are handled. For example, a request for a database connection might be handled on any one of four different servers in a cluster. Deciding which server handles the request depends on the load-balancing logic. Availability. Refers to the quality of the system's response to process a request, denoting the amount of time that an application or system is available to perform work. Availability is typically measured in percentage uptime. High Availability. Is defined as support to backup services in case of a system or node failure. A system with high availability uses load balancing so that all requests are routed to one or more preferred servers until one of the servers becomes unavailable, and then routed to another server that is still operational.
10
Failover. Is the ability for a request that is being serviced, to have a high availability switchover to another node without disruption of the service. A successful failover means to transfer the service without any intervention by client or manual process. Scalability. Denotes capability of a system to scale up to provide sustained system availability and throughput, without unacceptable degradation under increase in load. The increase in load could be due to increase in the number of users accessing the application, or in the volume or size of user transactions or actions. Reliability. Is the ability of a system to perform its required functions and sustain a requested service under stated conditions for a specified period of time. Horizontal Cluster. This refers to multiple installations of SBM, each hosting a cluster node, running in the same physical machine or multiple machines, but are part of the same cluster and runs homogeneously. Vertical Cluster. This refers to a single installation of SBM hosting multple cluster nodes, running in the same physical machine. A vertical cluster could be part of a horizontal cluster that is, a cluster having three nodes could have two nodes running in the same machine (vertical cluster) and the third node in a different machine (horizontal cluster). For more definitions, see the Glossary on page 80 of this guide. For a fuller listing of SBM terminology, see the Terminology Guide in the SBM_HOME\docs folder.
Extent of high availability needs Users, and the volume of the user transactions Network design and capacity Database design and dependencies Systems management
If the load requirements are unpredictable, more servers are needed in the configuration to ensure high availability and good performance at all times. On the other hand, if the risk of failure is high, along with the possibility of multiple failures, then more servers are required to be available for failover.
Clustering Guide
11
Clustering Guide
12
Chapter 2
Clustering in SBM
This chapter lists the new Clustering functions introduced in SBM 6.5 SP2, and describes the basic Clustering features.
Whats New
SBM provides the following new features for clustering:
Support for High Availability JMS with Pramati Cluster Ability to refresh configuration parameters across all nodes Support for Vertical Clustering
Clustering Features
SBM continues to support the following clustering features:
XML based cluster configuring. For more information, refer to Editing the Cluster Configuration File on page 32.
For more information, refer to Configuring Pramati Application Server on page 43, Configuring WebLogic Application Server on page 52, and Configuring WebSphere Application Server on page 64.
NOTE: On cluster, SBM Server as a Windows service is not supported. 13
Load Balancing
Large-scale business solutions require that the system is scalable to meet the performance and throughput demands with increase in workload. SBM employs load balancing to meet these demands. In Figure 2, the client sees the cluster as a single SBM setup. All three nodes have a common database. The requests sent by the client are shared by the three nodes. In case of BizLogic server load balancing, a group of application server clusters work as a single system sharing the load.
Note: BizLogic/BizPulse Web Services API and Process Repository are not supported in the clustered environment.
Failover
Large organizations focus on high availability of their systems, where the system must be configurable to support backup services in case of infrastructure component failure. The setup shown in Figure 2 represents three nodes, where only one node is always active. When the active node fails, one of the remaining nodes takes over, assuring failover. BizPulse is clustered for failover to ensure high availability.
Clustering Guide 14 Clustering SBM Servers
Vertical Clustering
SBM also provides support for vertical clustering, where a single SBM installation may host more than one node. This is achieved transparently and without any parameter or configuration changes. Log information for individual nodes is recorded in separate log files. Each component log file name would be appended with _<nodeName>. For example, if the nodes in vertical cluster are node1 and node2 then, bizlogic.log would be bizlogic_node1.log and bizlogic_node2.log.
Clustering Guide
15
In case of Balanced Scorecard Designer, the actual design is carried out from the BPM Portal, and the new application may get saved in any one node in the cluster. You must ensure that the physical copies of the application are available in all the nodes before installing the application. When you actually install the new Balanced Scorecard application, you need to copy the installed application from the node where it got installed, to all the other nodes in the cluster, similar to other BizLogic/ BizSolo applications. Similarly, publishing applications/worksteps as web services generates WSDL files on one of the nodes. You need to copy these files to all nodes to make web services run in the clustered environment.
Clustering Guide
16
Currently, Savvion offers nine Managed Adapters: J-Integra Excel Adapter, Database Adapter, iWay Adapter, File Adapter, Email Adapter, FTP Adapter , JMS Adapter, QuickRules Adapter and Web Service Adapter. When using a managed adapter in a clustered environment, you must ensure that all the files required for the work of the adapter are present in all the cluster nodes. These issues are described for the various managed adapters in the following section.
A J-Integra Excel Adapter requires a temporary directory to store its work files. You can specify this
directory through the Excel Adapter Configurator (for more information, see Chapter 3 in the Managed Adapters Guide). This temporary directory must be present on all the cluster nodes.
If you configured the J-Integra Excel Adapter to use a static Workbook file (as opposed to obtaining
the Workbook from a Document dataslot), make sure the Workbook is present at the same location on all the cluster nodes.
For iWay Adapters, the SBM_HOME\Schema directory and the Dictionary file must be available on
each of the cluster nodes at the same location.
The adapter jar and other jar files that the iWay Adapter needs, must be in the server classpath for
each node in the cluster. For example, if you are using the iWay RDBMS Adapter, include the iwrdbms.jar and any related database drivers in the server classpath on all the nodes.
Clustering Guide
17
Chapter 3
The Cluster Configuration Tool is a quick way of configuring the SBM cluster by providing the required system details and running the cluster setup directly from the Tool after it validates the settings. The cluster configuration is specified for the EJB cluster, JMS cluster, Web cluster, or any combination of these. The Cluster Configuration Tool works in environments that support graphics, and procedures for using this tool are covered in this chapter. For environments that do not support graphics, the file sbm-cluster.xml can be edited manually as described in Editing the Cluster Configuration File on page 32. The remaining chapters of this guide describe procedures for individual application servers providing advanced information, and you may require to refer to them.
18
In the left pane, the top node SBM Cluster Configuration displays the application server details.The Cluster Configuration Tool automatically detects and displays the application server and its relevant details. For example, the above figure shows the application server as pramati, and displays its version and protocol. If the application server is WebLogic, the tool will additionally display WebLogic domain and location. The Cluster Configuration Tool has the following tabs:
Setup: This tab shows the configuration panel, Cluster panel, and the Server panel. These are
described in detail in the following sections.
View XML: This tab displays the generated XML file after the setup configuration is verified. Run: This tab displays the log messages when the cluster is being deployed.
Depending on the selection in the left pane, the right pane displays different panels - Configuration panel, Cluster panel, and Server panel.
Clustering Guide
19
Configuration Panel
In the left pane, click the Clusters node or the Standalone Servers node. The Configuration panel is displayed in the right pane as shown in the following figure.
Figure 4 Configuration Panel
The Configuration panel displays links to the clusters or servers available as an alternative way of navigation.
Cluster Panel
In the left pane, click the Clusters node, and then click any cluster. Click any existing node, or click Add Node to display the Cluster panel in the right pane as shown in the following figure.
Figure 5 Cluster Panel
Server Panel
In the left pane, click on the Standalone Servers node, and then click any server. The Server panel is displayed in the right pane as shown in the following figure.
Figure 6 Server Panel
The Server panel displays information about the selected server. Note that EJB, JMS and Portal (Web) servers can be standalone or in cluster. If you have chosen a cluster for any of them, then the standalone server is disabled.
Note: The document server is required in case of Web cluster.
EJB Cluster
In your SBM setup, you can have either a single EJB server, or a cluster of EJB servers. To set up the EJB cluster, carry out the following steps: 1. In the left pane, under the node SBM Cluster Configuration, click Clusters.
Clustering Guide
21
2. Under the Clusters node, click EJB to display the EJB cluster setup as shown in the following figure.
Figure 7 Setting EJB Cluster
3. In the Cluster panel, click Add Node to display the node information as shown in the following figure.
Figure 8 EJB Cluster- Adding node
4. Specify Name for the node, which should be unique. We recommend that you do not use the character . in the node name to avoid any cluster start problems. 5. Specify the IP Address, HTTP Port and JNDI Port on which the node listens. 6. Enter the location of SBM Home, which is the installation directory of SBM.
Clustering Guide
22
7. Click Add to add the node to the EJB cluster. Click Add & Continue to add more nodes. Click Cancel to go back to the previous screen. 8. Click Save to save the configuration.
JMS Cluster
In your SBM setup, you can have either a single JMS server, or a cluster of JMS servers. To set up the JMS cluster, carry out the following steps: 1. In the left pane, under the node SBM Cluster Configuration, click Clusters. 2. Under the Clusters node, click JMS to display the JMS cluster setup as shown in the following figure.
Figure 9 Setting JMS Cluster
Clustering Guide
23
3. In the Cluster panel, click Add Node to display node information as shown in the following figure.
Figure 10 JMS Cluster- Adding node
4. Specify Name for the node, which should be unique. We recommend that you do not use the character . in the node name to avoid any cluster start problems. 5. Specify the IP Address and JNDI Port on which the node listens. 6. Enter the location of SBM Home, which is the installation directory of SBM. 7. Name the Target Node. 8. Click Add to add the node to the JMS cluster. Click Add & Continue to add more nodes. Click Cancel to go back to the previous screen. 9. Click Save to save the configuration.
Web Cluster
In your SBM setup, you can have either a single Portal server, or a Web cluster of Portal servers. To set up the Web cluster carry out the following steps: 1. In the left pane, under the node SBM Cluster Configuration, click Clusters.
Clustering Guide
24
2. Under the Clusters node, click Web to display the Web cluster setup as shown in the following figure.
Figure 11 Setting Web Cluster
3. In the Cluster panel, click Add Node to display node information as shown in the following figure.
Figure 12 Web Cluster- Adding node
4. Specify Name for the node, which should be unique. We recommend that you do not use the character . in the node name to avoid any cluster start problems. 5. Specify the IP Address, HTTP Port and JNDI Port on which the node listens. 6. Enter the location of SBM Home, which is the installation directory of SBM. 7. Click Add to add the node to the Web cluster. Click Add & Continue to add more nodes. Click Cancel to go back to the previous screen.
Clustering Guide 25 Setting Up the Cluster
Standalone Servers
In your SBM setup, you can have either clusters of EJB/JMS/Portal servers, or standalone single servers. The following procedure describes the configuration of the standalone servers. 1. In the left pane, under the node SBM Cluster Configuration, click Standalone Servers. 2. In the Server panel, click the server you want to configure, to display its information as shown in the following figure.
Figure 13 Setting Standalone Servers
The names of the Standalone EJB and Portal servers must be ejbServer and
Clustering Guide
26
Proxy Server
A proxy server is required with a Web cluster. When a Web cluster is set up, the external clients send their request to the proxy server. Then, it is the proxy server that routes the request to one of the servers in the Web cluster. Therefore the proxy server configuration is essential with the Web cluster. The proxy server is provided as a plugin by the application server vendor and is different for different Web servers. For installation of plugin in Web servers, refer to the vendor specific documentation. If you are using WebLogic as httpserver, then it already includes the plugin and the cluster setup utility configures the proxy server. A proxy server need not always be a new server instance. In case of WebSphere, the IBM HTTP Server is used as a proxy server. In case of Pramati and WebLogic, a new server instance is created. The following procedure describes the configuration of the Proxy server. 1. In the left pane, under the node SBM Cluster Configuration, click Standalone Servers. 2. In the Server panel, click Proxy Server to display its information as shown in the following figure.
Figure 14 Setting Proxy Server
3. Specify Name for the server, which should be unique. 4. Specify the IP Address and HTTP Port on which the server listens. 5. Click Save to save the configuration.
Clustering Guide
27
Document Server
The Document server in SBM that deals with the documents related to the Document dataslots also needs to be configured. In case of the Web cluster, the documents must be stored on a shared drive so that all nodes can access this location. The document location properties are also modified in the files sbm.conf and designer.conf when the modifyprop.bat utility is run.
Note: The document server is required in case of Web cluster.
The following procedure describes the configuration of the Document server. 1. In the left pane, under the node SBM Cluster Configuration, click Standalone Servers. 2. In the Server panel, click Document Server to display its information as shown in the following figure.
Figure 15 Setting Document Server
3. Specify Name for the server, which should be unique. 4. Specify the URL and the Document Location of the server. Click Browse to search the location. 5. Click Save to save the configuration. URL for the Document Server The URL for the Document server depends on whether a single Portal server has been configured or a Web cluster. For a single Portal server, the Document server URL is the same as the URL of the single Portal server.
Clustering Guide
28
For a Web cluster, if there is a drive controller in Windows using RAID controller or a shared drive on UNIX, then the document server URL can be specified as the proxy server URL. This will provide failover. SBM validates if the Document server URL is same as the Proxy server URL and also if the document location is specified for the Document server. If validation fails, an exception is thrown. The validation can be disabled by setting the system property sbm.cluster.extdocserver to True in the file SBM_HOME\cluster\conf\sbmclusterprops.xml.
Clustering Guide
29
Clustering Guide
30
3. Make changes as necessary. 4. Click Update to save the changes. Click Cancel to go back to the previous screen. 5. In Figure 16, click Delete to remove the node from the cluster. When the last node in a cluster is deleted, the cluster no longer exists.
Clustering Guide
31
Chapter 4
The Cluster Configuration Tool works in environments that support graphics. If you are using an environment that supports graphics, we recommend you use the Cluster Configuration Tool for configuration as described in Using the Cluster Configuration Tool on page 18. For environments that do not support graphics, you can manually edit the cluster configuration file sbm-cluster.xml to provide the cluster details. The remaining chapters of this guide describe procedures for individual application servers providing advanced information, and you may require to refer to them.
Note: You can specify a custom file name to the cluster configuration file other than sbm-cluster.xml while ensuring that the file conforms to sbm-cluster.dtd. If no custom file name is specified, the default name sbm-cluster.xml is used.
The cluster configuration is specified in an XML file placed in SBM_HOME\cluster\conf. Using this configuration file, you can specify the EJB cluster, JMS cluster, Web cluster, or any combination of these. Clustering in SBM has the following common prerequisites:
Each node should have a static IP address. Standalone EJB and Portal servers must not be running. Machines and port numbers must be selected after proper planning.
For clustering with Pramati application servers, refer to Configuring Pramati Application Server on page 43. For clustering with WebLogic application servers, refer to Configuring WebLogic Application Server on page 52. For clustering with WebSphere application servers, refer to Configuring WebSphere Application Server on page 64.
32
General Tags
The configuration file uses the following general tags. <cluster-configuration> This is the main tag in the configuration file. Attributes:
appserver: This is the name of the application server. For example, Pramati. appversion: This is the version of the application server.
<domain-name> This is an optional tag for Pramati. This is a mandatory tag for WebLogic. Specify the domain name with which SBM is installed under WebLogic in the file SBM_HOME\cluster\conf\sbm-cluster.xml.
Note: The domain name is case sensitive.
Along with the domain-name, you need to specify the location, which is the absolute path of the directory under which your SBM is installed in WebLogic. For example, <domain-name location="C:\bea\user_projects\domains\sbm50">sbm50 </domain-name> <protocol> A <cluster-configuration> has a protocol that is required to communicate to the application server and create the URLs for lookup. For example t3:// for WebLogic. The protocol can be overridden for an individual <cluster> by defining it within the cluster. For example, in case of a Web cluster which uses HTTPS connection, it will look as follows: <cluster name="WebCluster" type="web"> <protocol>https://</protocol> ... This tag is optional. <synchronize-time> A <cluster-configuration> has a <synchronize-time> tag that is required for synchronizing the time of all the node machines in the cluster. In this tag, the node attribute should have the name of the node, which is taken as the reference and all the other nodes time is synchronized with the reference node.
Clustering Guide
33
The cluster nodes are synchronized with the target node time only during the cluster setup. Any exception during the synchronization of node time will be ignored and the cluster setup will continue.
Note: This tag is supported only for Windows and Solaris operating systems.
Any errors during synchronization are ignored and the cluster setup is continued. Note that the utility does not log any errors outside its control. You can also use third-party tools for synchronizing time. <authentication> A <cluster-configuration> has an <authentication> tag that provides the user name and password to connect to the application server for the cluster setup. The user name and password should have administration rights. For example, system/wlsysadmin can be used for WebLogic and root/pramati can be used for Pramati. <multicast> IP multicast is a simple broadcast technology that enables multiple applications to "subscribe" to a given IP address and port number and listen for messages. A multicast address is an IP address in the range from 224.0.0.0 to 239.255.255.255. The cluster requires a <multicast> tag for WebLogic that corresponds to the multicast address. It has a port attribute that has the multicast port number.
Note: The multicast address or port should be different for each application cluster. An Application cluster may include EJB cluster, JMS cluster and Web cluster types. This is important if more than one application cluster runs in the same intranet.
<cluster> A <cluster-configuration> can have multiple <cluster> tags corresponding to each cluster to be set up, for example, one for EJB cluster, one for JMS cluster and one for Web cluster. There could be only one cluster defined for a type. Each cluster will have more than one node defined under the <nodes> tag. The <nodes> tag has as many <node> tags as the number of nodes in the cluster. Attributes:
type: This is the type of cluster. Valid types are ejb, jms and web.
<node> This represents a node in the cluster. It has a single attribute:
name: The unique name of the node. It may be same as the host name, or different from the host
name.
Clustering Guide
34
<address>: This is the IP address of the node or its host name. <jndi-port>: This is the JNDI port on which the node listens. Normally the naming service runs on
this port. This tag is valid for EJB, JMS and WEB nodes.
<http-port>: This is the HTTP port on which the node listens. This tag is valid for EJB, WEB and
PROXY server.
<sbm-home>: This is the location where SBM is installed in the cluster nodes.
The cluster utility creates a server instance for each node with the server name as the node name. For example, in case of WebLogic, two servers will be created with the name ejb1 and ejb2 for the following configuration: <cluster name="ejb_cluster" type="ejb" > <nodes> <node name="ejb1"> . . . </node> <node name="ejb2"> . . . </node> </nodes> </cluster>
Note:
We recommend that you do not use the character . in the node name to avoid any cluster start problems.
<url>: The database URL. <driver>: The fully qualified class name of the database driver. <user>: Database user name.
Clustering Guide 35 Understanding the Tags
<password>: The password for the above user. <provider>: The database used: for example, oracle, db2. <databasename>: Name of the database: for example, db2. <connection-pool-name>: Name of the connection pool defined in the appserver, required only for
WebLogic JMS cluster database persistence. For application servers like Pramati, if all the clusters (EJB, JMS and Web) use the same database, then similar persistence information needs to be replicated for each cluster. But keeping the persistence with each cluster provides the flexibility to have a different database or the same database but different users for different clusters. For example, the JMS and EJB clusters might use different databases.
For JMS server: jmsServer For Portal server: portalServer For Admin server: adminServer
Note that adminServer, ejbServer, and portalServer should be valid active server instances which should already exist before the cluster setup is started. Also note that jmsServer and documentServer are the logical references pointing to a valid active WebLogic server instance. The proxyServer is a server instance created by the cluster utility. The port specified for proxyServer should be unique; that is, it should not conflict with any of the running servers ports in the cluster.
Note: For a particular type (JMS/Web) there can either be a single server or a cluster, however, both single server and cluster for the same type cannot exist together.
Order of Tags
Ensure that the order of tags in the file sbm-cluster.xml is the same as defined in the file sbm-cluster.dtd. If the sequence is not the same as defined in the DTD file, then while starting the cluster setup, you will get the following exception: java.io.IOException: IO error: The content of element type "cluster-configuration" must match "(domain-name?,protocol?,synchronize-time?,authentication?,multicas t?,cluster+,adminserver?,proxyserver?,documentserver?,ejbserver?,po rtalserver?,jmsserver?,deploy?)"
Clustering Guide 36 Understanding the Tags
The order of the elements as defined in the DTD file is as follows: <!ELEMENT cluster-configuration (domain-name?, protocol?, synchronize-time?, authentication?, multicast?, cluster+, adminserver?, proxyserver?, documentserver?, ejbserver?,portalserver?, jmsserver?, deploy?)> If you want to set up both EJB and JMS clusters, then definitions of both the clusters should come at the top after the multicast element.
EJB Setup
In your SBM setup, you can have either a single EJB server, or a cluster of EJB servers.
Note: If a single EJB server is specified, then it uses the embedded JMS server, and an additional JMS server should not be created. SBM validates that the EJB server and JMS server are identical in name, address, and jndi-port number.
Single EJB Server Single EJB server is specified with the help of the <ejbserver> element. The name attribute holds the unique name of the EJB server. The IP address and the port on which the server listens are specified with the help of the <address> and <jndi-port> elements under the <ejbserver>.
Note: If a single EJB server is specified, then an EJB cluster cannot be specified and vice versa.
EJB Cluster For an EJB cluster, you need to provide the following information:
The name attribute of <cluster> element should contain a unique name for the EJB cluster. The type attribute of the <cluster> element should be "ejb". For each ejb node in the cluster, specify a <node> element under the <nodes> in <cluster>. The name attribute of the <node> will contain the unique name of the ejb node. The IP address of
the EJB node and the port on which it listens are specified with the <address> and <jndi-port> elements under the <node>.
Note: For any validation related to the EJB cluster, SBM refers to the attribute if-cluster-type in the file SBM_HOME\cluster\conf\sbm-properties.xml.
Clustering Guide
37
JMS Setup
In your SBM setup, you can have either a single JMS server, or a cluster of JMS servers.
Note: If a single EJB server is specified, then it uses the embedded JMS server, and an additional JMS server should not be created. SBM validates that the EJB server and JMS server are identical in name, address, and jndi-port number.
Single JMS Server Single JMS server is specified with the help of the <jmsserver> element. The name attribute holds the unique name of the JMS server. The IP address and the port on which the server listens are specified with the help of the <address> and <jndi-port> elements under the <jmsserver>.
Note: If a single JMS server is specified, then a JMS cluster cannot be specified and vice versa.
JMS Cluster For a JMS cluster, you need to provide the following information:
The name attribute of <cluster> element should contain a unique name for the JMS cluster. The type attribute of the <cluster> element should be "jms". The store attribute of the <cluster> element should be "db" for database persistence or "file" for file
persistence.
For each JMS server in the cluster, specify a <node> element under the <nodes> in <cluster>. The name attribute of the <node> will contain the unique name of the JMS server. The IP address of
the JMS server and the port on which it listens have to be specified with the <address> and <jndi-port> elements under the <node>.
The target of the JMS server specified under <target-node> should be any server instance. The tag
<target-node> is required only for WebLogic JMS cluster.
If file persistence is used, then a <file-store> is to be specified under each <node> that will contain
the absolute path to the file that is to be used as the store.
If database persistence is used, then a single <persistence> tag needs to be defined under the
<cluster>. JMS Cluster Persistence The persistence properties are defined with the help of <persistence> element under the <cluster> element. You need to provide the values for the provider, url, driver, user name, password and name of the connection pool such as SBMCommonDBPool.
Clustering Guide
38
Portal Setup
In your SBM setup, you can have either a single Portal server, or a Web cluster of Portal servers. Single Portal Server A single Portal server is specified with the help of the <portalserver> element. The name attribute holds the unique name of the Portal server. The IP address and the port on which the server listens are specified with the help of the <address> and <http-port> elements under the <portalserver>.
Note: If a single Portal server is specified, then a Web cluster cannot be specified and vice versa.
Web Cluster For a Web cluster, you need to provide the following information:
The name attribute of <cluster> element should hold a unique name for the Web cluster. The type attribute of the <cluster> element should be "web". For each Web server in the cluster, specify a <node> element under the <nodes> in <cluster>. The name attribute of the <node> will contain the unique name of the Web server. The IP address of
the Web server and the port on which it listens have to be specified with the <address> and <http-port> elements under the <node>.
Proxy Server
A proxy server is required with a Web cluster. When a Web cluster is set up, the external clients have to send their request to the proxy server only. Then, it is the proxy server that routes the request to one of the servers in the Web cluster. Therefore the proxy server is essential with the Web cluster. A proxy server is specified with the help of the <proxyserver> element. The name attribute holds the unique name of the proxy server. The IP address and the port on which the server listens, are specified with the help of the <address> and <http-port> elements under the <proxyserver>. The proxy server is provided as a plugin by the appserver vendor and is different for different Web servers. If you are using WebLogic as httpserver, then it already includes the plugin and the cluster setup utility configures the proxy server. A proxy server need not be a new server instance. In case of WebSphere, the IBM HTTP Server is used as a proxy server. In case of Pramati and WebLogic, a new server instance is created. For installation of plugin in Web servers, refer to the vendor specific documentation. The values for the <address> tag under <proxyserver> tag should be of the server where you installed the plugin.
Document Server
The Document server in SBM that deals with the documents related to the Document dataslots also needs to be configured in the configuration file. This can be done with the help of the <documentserver> element under <cluster-configuration>. The <documentserver> element has a name attribute that is the unique name of the document server. It consists of the elements: <doc-location> and <url>.
Clustering Guide
39
The <documentserver> element is mandatory for the Web cluster. The documents must be stored on a shared drive so that all nodes can access this location. The <doc-location> element specifies the exact location of the documents directory used for storing the document dataslot contents. This element is optional. The document location properties can also be modified in the files sbm.conf and designer.conf when the modifyprop.bat | sh utility is run. In this case, the <doc-location> element need not be specified.
Note: The element <documentserver> is required with the Web cluster.
URL for the Document Server The URL for the Document server depends on whether a single Portal server has been configured or a Web cluster. For a single Portal server, the Document server URL is the same as the URL of the single Portal server. For a Web cluster, if there is a drive controller in Windows using RAID controller or a shared drive on UNIX, then the document server URL can be specified as the proxy server URL. This will provide failover. SBM validates if the Document server URL is same as the Proxy server URL and also if the document location is specified for the Document server. If validation fails, an exception is thrown. The validation can be disabled by setting the system property sbm.cluster.extdocserver to True in the file SBM_HOME\cluster\conf\sbmclusterprops.xml.
EJB Cluster
For an EJB cluster, the following tags should be included in the file sbm-cluster.xml:
<cluster name="sbm_cluster" type="ejb" create="true"> <adminserver name="wladmin"> <documentserver name="docServer"> <portalserver name="portalServer"> <jmsserver name="sys97">
Clustering Guide
40
JMS Cluster
Once the EJB cluster setup is done, the XML file should have the following tags for the JMS cluster setup:
<cluster name="web_cluster" type="web" create="true"> <adminserver name="wladmin"> <documentserver name="docServer"> <proxyserver name="proxy" type="pramati">
2. Start the application server node and verify that it is started properly. 3. From SBM_HOME\bin, run the command: addBizLogicNodes -u <user> -p <password> -nc <count> Where count is the number of new nodes to be added. This will synchronize the BizLogic server on the new node. 4. For BizPulse, when you add a new node, you need to start BizPulse monitoring on that node for failover. Refer to Starting and Stopping BizPulse Monitoring on page 79 for related procedures.
Note: BizPulse runs in a failover configuration and need not be re-synchronized.
Cluster configuration file sbm-cluster.xml is read from the file system by default. If the property
propertyfilestorage is set to True, the file will be read from the database.
If the database storage is enabled, the following files must be present in database:
o o sbm-cluster.xml sbm-properties.xml
For WebLogic Application Server, sbm-resources-weblogic.xml must also be present in the database.
You can check in and check out the cluster files using the script
SBM_HOME\cluster\bin\DBFileManager.
Clustering Guide
42
Chapter 5
This chapter describes the details of Pramati application server required for manual editing of the file sbm-cluster.xml. For basic reference information on manual cluster configuration, refer to the chapter Editing the Cluster Configuration File on page 32. If you plan to use the Cluster Configuration Tool instead, refer to the chapter Using the Cluster Configuration Tool on page 18. You can also use the file SBM_HOME\cluster\docs\readme_pramati.txt for quick reference.
Note: Portal cluster is not supported in vertical clustering using Pramati application server. When you install a BizLogic/BizSolo application in a clustered environment, it creates the application files on any one node in the cluster. You need to copy them from the node where it got installed, to all the other nodes in the cluster. You can avoid copying the files manually by using the BPM Studio. Refer to Using BPM Studio in Clustered Environment on page 16 for more information.
Before Clustering
Before clustering, ensure that:
Pramati Admin Service must be running on all intended nodes before starting the cluster setup. This
can be done by running the following command: SBM_HOME\pramati\server\bin\runstartupsvc.bat
43
On each node of the cluster, run the utility SBM_HOME\bin\setupSBM.bat | sh before starting the
cluster.
Verify that the table with name of the EJB cluster does not already exist in the database specified in
the persistence properties. For example, if the cluster name is sbmejbCluster, a table by this name is already created in the database specified in the <persistence> section of this cluster. If such a table already exists, then you need to delete it manually before running the utility.
Verify that the table with name of the JMS cluster does not already exist in the database specified in
the persistence properties. For example, if the cluster name is sbmjmsCluster, a table by this name is already created in the database specified in the <persistence> section of this cluster. If such a table already exists, then you need to delete it manually before running the utility.
Using Ports
While specifying the port numbers in the cluster configuration file, note the following points:
All EJB nodes should have the same port number. For example, if you choose port 9295, then
<cluster name="sbmejbCluster" type="ejb"> <nodes> <node name="sys91"> <address>10.1.32.129</address> <jndi-port>9295</jndi-port> <source-node>ejbServer</source-node> </node>
Clustering Guide
44
Before Clustering
<node name="sys74"> <address>10.1.32.112</address> <jndi-port>9295</jndi-port> </node> </nodes> The tag <http-port> can also be specified to access the HTTP port (if available) on this EJB node. This tag is optional.
All Web nodes should have the same port number. For example, if you choose port 8181,then since
Web cluster uses existing Portal servers, the port numbers will be the port numbers of the individual port numbers of the Portal servers. <cluster name="sbmwebCluster" type="web" create="false"> <nodes> <node name="sys91WA"> <address>10.1.32.129</address> <jndi-port>9191</jndi-port> <http-port>8181</http-port> </node> <node name="sys80WA"> <address>10.1.32.118</address> <jndi-port>9191</jndi-port> <http-port>8181</http-port> </node> </nodes> </cluster> The tag <jndi-port> needs to be specified for portal server nodes for query service to access the JNDI resources (JDBC, etc.) on the portal server nodes. This port will be different than <http-port> for Pramati server.
In case of vertical clustering, nodes are on the same machine, and port numbers need to be different.
Configuring Clusters
Note the following points:
The <source-node> tag is mandatory. Specify the "ejbServer" node as source node. This is used to
set the required SBM configuration on the cluster nodes. "ejbServer" is the default node of an SBM standalone installation.
The <persistence> tag is mandatory for EJB cluster. If the Web cluster option is used, then proxy server should be created on the same node as the
machine where the utility setupcluster.bat | sh is being run.
The standalone JMS server name (for example, sys91jms) must be different from any other node
name.
Clustering Guide
45
Configuring Clusters
Execute setupcluster.bat|sh -cmd in SBM_HOME\cluster\bin on one of the cluster nodes. Setting up High Availability JMS Cluster
When a High Availability (HA) JMS is set up with a Pramati Application Server, one of the JMS nodes is in active state and all other JMS nodes are in passive state. When the active JMS node fails, one of the other passive nodes becomes active. The <nodes> tag contains the list of JMS nodes for the cluster. The first <node> element must contain the <node name>, <address>, <jndi port> and <sbm home> tags. When you set up the HA JMS using the Cluster Configuration Tool, it enters the values for these tags. To set up the HA JMS with Pramati cluster, carry out the following steps: 1. Set up the JMS cluster using the Cluster Configuration Tool as explained in the Chapter 3, Using the Cluster Configuration Tool. 2. Similarly, add EJB and Portal nodes as desired. 3. Click the Run tab and click Run Setup to run the cluster. 4. From the machine on which the Cluster Configuration Tool is run, open the jms-config.xml file from <SBM_HOME>\pramati\server\nodes\<JMS_NODE_NAME>\config folder. 5. Locate the <persistent-store> tag. 6. Change the <persistence-type> tag value from file-store to db-store.
Clustering Guide 46 Setting Up the Cluster
7. Save this file. 8. From the machine on which the Cluster Configuration Tool is run, execute following command from <SBM_HOME>\pramati\server\bin folder: startJMSServer.bat -node <JMS_NODE_NAME> -uploadconfig This will save the changes to the database. 9. Stop the JMS Server by using the shutdown command in JMS Clusters shell. 10. Continue with the rest of the cluster setup procedure explained in the following sections.
Clustering Guide
47
bizpulse.server.cluster=on
This parameter should be set to on. This indicates that there are more than one nodes working in cluster mode.
bizpulse.server.cluster.monitor=on
Failover is achieved by monitoring the active BizPulse server, and initiating the failover action in case the server fails. Therefore, the word "Monitoring" is used here almost synonymously as "Failover". This parameter should be set to on to enable failover.
bizpulse.server.cluster.monitor.interval=30
When BizPulse monitoring is on, whenever BizPulse is started, a monitoring process will start on each of the nodes. BizPulse server starts up on any one of the nodes. On the node, where the BizPulse server is active, monitoring process checks the state of the BizPulse at the interval specified by bizpulse.server.cluster.monitor.interval in seconds. In this particular example, the monitoring takes place every 30 seconds, and the process writes the last BizPulse alive time to the database. On all the other nodes, monitoring process checks and compares the last BizPulse alive time with the current time.
bizpulse.server.cluster.failover.interval=120
When the difference between the last BizPulse alive time and the current time exceeds the specified bizpulse.server.cluster.failover.interval, the monitoring process will try to start the BizPulse server on a node other than the one on which BizPulse was last active.
Clustering Guide
48
2. If the Web cluster is not set up, and there is a single Portal server, start the Portal server with the command: SBM_HOME\pramati\server\bin\startPortalServer.bat|sh. 3. Start the JMS server with the command: SBM_HOME\pramati\server\bin\startJMSServer.bat|sh <JmsServer> In this example, the JmsServer name is sys91jms.
Note: You need to run this command on all the nodes individually. Always start JMS server / cluster before starting EJB server / cluster.
4. If the EJB cluster is set up, start the EJB servers of the cluster with the command: SBM_HOME\pramati\server\bin\startEjbServer.bat|sh [nodename] You must run the command for each node. In this example, the nodename is sys91 and sys74. 5. If the EJB cluster is not set up, and there is a single EJB server, start it with the command: startEJBServer
For more information on BizLogic server commands, refer to BizLogic Server Commands on page 76. Starting the BizPulse Server Go to SBM_HOME\bin and start the BizPulse server as follows: 1. If the EJB Cluster is set up, then start with the command startBizPulseServer -nodes <nodes in cluster> 2. If the EJB Cluster is not set up, and there is a single EJB server, then start with the command startBizPulseServer For more information on BizPulse server commands, refer to BizPulse Server Commands on page 78.
Performance Tuning
To improve the performance, you can carry out the following tuning: 1. In the out-of-the-box installation, there are two copies of web-config.xml; namely, web-config.xml.dev and web-config.xml.prod under SBM_HOME\pramati\server\nodes\portalServer\config. The web-config.xml.prod file is tuned for a production environment with the update-check-interval-seconds value set to -1. This disables the check the server performs for changes in JSP files. In the production environment, rename this file to web-config.xml and restart the server.
Note: Once the update-check-interval-seconds value is set to -1, the Portal server will no longer compile JSPs. If you want to deploy new JSPs or need the server to recompile modified JSPs, the update-check-interval-seconds value should be a positive value.
2. The values of keep-alive-timeout (millis) and socket-idle-time-out (millis) should be changed to a high value in the web-config.xml of the backend nodes, (preferably to more than five minutes for keep-alive-timeout and 10-60 seconds for idle-timeout). This ensures that the socket connections between the Loadbalancer and the backend nodes does not get timed out for a long period of time, and hence does not result in unnecessary, and frequent creation of sockets. For each of the backend nodes in SBM_HOME\pramati\server\nodes\portalServer\config\web-config.xml change the idle-timeout-millis to 10000. <idle-timeout-millis>10000</idle-timeout-millis> <keepalive-timeout-millis>30001</keepalive-timeout-millis>
Clustering Guide
50
Performance Tuning
3. It is okay to turn off the DOS (Denial of Service attack) protection on the backend nodes in a cluster. This can be done by setting the denial-of-service enabled to false. Since the DOS protection is turned on in the proxyServer/loadbalancer by default, turning it off in the backend nodes will avoid DOS attack checks from happening twice (in the loadbalancer and the backend node), thereby improving performance. a. For the proxy node, in SBM_HOME\pramati\server\nodes\proxy\config\web-lbconfig.xml, in the nodes section, make sure that the socket-idle-timeout-millis="10000" for both the nodes. For example, <node type="J2EE" host="jerusalem.savvion.com" web-port="8181" name="web_jerusalem"> <naming-port /> <socket-pool enabled="true" min="20" max="50" pool-idle-time-out-seconds="1000" socket-idle-timeout-millis="10000" /> </node> <node type="J2EE" host="kuba.savvion.com" web-port="8181" name="web_kuba"> <naming-port /> <socket-pool enabled="true" min="20" max="50" pool-idle-time-out-seconds="1000" socket-idle-timeout-millis="10000"/> </node> b. The proxyServer\loadbalancer node is configured for a session timeout of 6 min by default. This can be updated to a desirable value by modifying the code as follows: <node-chooser name="session-stickiness-node-chooser" class="com.pramati.web.lb.nodechooser.SessionStickinessNodeChoo ser" enabled="true"> <props time-out-minutes="6" cleanup-interval-minutes="2" /> </node-chooser> to <node-chooser name="session-stickiness-node-chooser" class="com.pramati.web.lb.nodechooser.SessionStickinessNodeChoo ser" enabled="true"> <props time-out-minutes="45" cleanup-interval-minutes="2" /> </node-chooser>
Note: The sample values used in this example should be changed to the appropriate values in your system.
Clustering Guide
51
Performance Tuning
Chapter 6
This chapter describes the details of WebLogic application server required for manual editing of the file sbm-cluster.xml. For basic reference information on manual cluster configuration, refer to the chapter Editing the Cluster Configuration File on page 32. If you plan to use the Cluster Configuration Tool instead, refer to the chapter Using the Cluster Configuration Tool on page 18. You can also use the file SBM_HOME\cluster\docs\readme_weblogic.txt for quick reference. Clustering in SBM uses the following terminology:
SBM_HOME: Refers to the installation directory of SBM. WL_HOME: Refers to the WebLogic installation directory WL_SBM_DOMAIN: Refers to the absolute location where SBM is installed under WebLogic. For
example, for WebLogic 8.1, C:\bea\user_projects\domains\sbm50.
Note: When you install a BizLogic/BizSolo application in a clustered environment, it creates the application files on any one node in the cluster. You need to copy them from the node where it got installed, to all the other nodes in the cluster. You can avoid copying the files manually by using the BPM Studio. Refer to Using BPM Studio in Clustered Environment on page 16 for more information.
52
Before Clustering
Before clustering, ensure that:
If a single JMS server is configured for the cluster, then this server should be targeted to the EJB
server that is at the same location as the Administration server. For example, if the Administration server is at node1 IP=10.1.32.129 and EJB nodes are 10.1.32.129 and 10.1.32.179, then the JMS server should always be targeted on 10.1.32.129.
Each node should be installed on a separate machine, and all nodes should use the same database
account.
Ensure that SBM is installed on all nodes before proceeding. You are required to use the same
directory for each node, and the same WebLogic domain name for each node.
Make sure that the Admin Server is running on one of the machines in the cluster before starting the
cluster setup.
On each node in the cluster, run the utility SBM_HOME\bin\setupSBM.bat |sh. For more
information, see the Installation Guide.
The jar and ear files included with default installation must not be clustered.
Using Ports
While specifying the port numbers in the cluster configuration file, note the following points:
All EJB nodes should have the same port number. For example, if you choose 18785, then
<cluster name="sbmejbCluster" type="ejb"> <nodes> <node name="sys97"> <address>192.168.0.28</address> <jndi-port>18785</jndi-port> </node> <node name="sys98"> <address>192.168.0.32</address> <jndi-port>18785</jndi-port>
Clustering Guide
53
Before Clustering
</node> </nodes> </cluster> The tag <http-port> can also be specified after <jndi-port> tag to access the HTTP port (if available) on this EJB node. This tag is optional.
All Web nodes should have the same port number. For example, if you choose 18795, then since the
Web cluster uses existing Portal servers, the port numbers will be the port numbers of the individual port numbers of the Portal servers. <cluster name="sbmwebCluster" type="web"> <nodes> <node name="sys97_web"> <address>192.168.0.28</address> <http-port>18795</http-port> </node> <node name="sys98_web"> <address>192.168.0.32</address> <http-port>18795</http-port> </node> </nodes> </cluster> The tag <jndi-port> can be specified for portal server nodes for query service to access the JNDI resources (JDBC, etc.) on the portal server nodes. This port is the same as the <http-port> for WebLogic server, and hence can be omitted.
All JMS nodes should have the same port number as EJB nodes. For example, 18785. Also, the IP
address and target of JMS nodes should be the same as the IP address and node name of the corresponding EJB nodes respectively. <cluster name="sbmjmsCluster" type="jms" store="db"> <nodes> <node name="sys97_jms"> <address>192.168.0.28</address> <jndi-port>18785</jndi-port> <target-node>sys97</target-node> </node> <node name="sys98_jms"> <address>192.168.0.32</address> <jndi-port>18785</jndi-port> <target-node>sys98</target-node> </node> </nodes>
NOTE: The tag <http-port> is not valid for JMS server or JMS cluster nodes.
Clustering Guide
54
Before Clustering
You must use the same port number specified while installing SBM. For example, 18791.
<adminserver name="adminServer"> <address>192.168.0.28</address> <jndi-port>18791</jndi-port> </adminserver>
In case of vertical clustering, (nodes on same machine), port numbers need to be different.
Configuring Clusters
Note the following points:
The <domain-name> tag is mandatory for WebLogic. Specify the domain name with which SBM is
installed under WebLogic in the file SBM_HOME\cluster\conf\sbm-cluster.xml. The domain name is case sensitive.
The location attribute is mandatory for WebLogic. Specify the absolute path of the directory under
which SBM is installed in WebLogic. For example, <domain-name location= "C:/bea/user_projects/domains/sbm50">sbm50 </domain-name>
If SBM is installed in different directories on different nodes during the cluster setup with
WebLogic, then the file config.xml on the Admin server needs to be changed manually. You need to first run the cluster setup utility, then shut down the Admin sever, and then edit the file as follows: The paths of the applications in the file config.xml are absolute by default. You need to make the paths relative to the domain directory. For example, change the path of the file bizlogic.ear from \u4\bea\user_projects\domains\SBM\applications\bizlogic.ear to .\applications\bizlogic.ear
If SBM is installed in different locations on the cluster nodes, and a single JMS server is to be
configured, then the property sbm-home needs to be added for all the nodes in the file sbm-cluster.xml. <node name="sys113"> <address>10.1.32.172</address> <jndi-port>18785</jndi-port> <sbm-home>D:/SBM_HOME</sbm-home> </node> Where, sbm-home is the location of SBM installation on that specific node.
Clustering Guide
55
Configuring Clusters
In case of a single JMS server, if the SBM_HOME property is different for at least one node, then
the utility sets a pre-configured JDBC store (BizLogic_JDBC_Store, BizPulse_JDBC_Store) for BizLogic and BizPulse. In case of JMS cluster, this property is ignored. By default, the JMS store for a standalone installation is a file on the local file system. If SBM is installed in different locations on different nodes, then the file store path of an assigned JMS server may not be valid on all the machines. The JMS server (and hence the EJB node) will not start in such case. Hence a JDBC store is set as the JMS store to avoid this problem.
</nodes> . . . </cluster> JMS Cluster with EJB Cluster If you configure the JMS Servers with an EJB Cluster, then the value of <target-node> should be set to the EJB Cluster node names. Similarly, the address and port of the JMS servers in the JMS Cluster should be the same as that of the EJB Servers in the EJB Cluster. The name of the JMS Servers in the JMS Cluster should not exceed 16 characters. An example of this is given below: <cluster name="ejb_cluster" type="ejb"> <nodes> <node name="sys97"> <address>192.168.0.28</address> <jndi-port>9000</jndi-port> </node> <node name="sys98"> <address>192.168.0.32</address> <jndi-port>9001</jndi-port> </node> </nodes> </cluster> <cluster name="sbmjmscluster" type="jms" store="db"> <nodes> <node name="jmsNode1"> <address>192.168.0.28</address> <jndi-port>9000</jndi-port> <target-node>sys97</target-node> </node> <node name="jmsNode2"> <address>192.168.0.32</address> <jndi-port>9001</jndi-port> <target-node>sys98</target-node> </node> </nodes> . . . </cluster>
Clustering Guide
57
Configuring Clusters
While the Portal server is made clusterable, an application with name WLProxyApp.war is uploaded by the cluster utility in the upload folder of the WL_SBM_DOMAIN\adminServer folder. This can be seen under the SBM_HOME\cluster\webapps folder.
Note: If you need to install a proxy server in other webservers/servlet engines, then it is required to download and install the WebLogic plugins. Refer to the WebLogic documentation for downloading, installing and configuring the proxy server for IIS and Netscapes iPlanet.
After the setup is complete, you need to shut down the Admin server and then restart it. If the server does not start properly, then the cluster setup remains incomplete, and results in a corrupted config.xml file. After the setup is completed, check the application server log file WL_SBM_DOMAIN\adminServer\logs\adminServer.log for any errors in setup. In some cases, the utility does not get any exception from the server and continues to show that the setup is okay. This may happen for various reasons such as an incorrect domain name.
Note: The cluster utility must be run from the same machine on which the WebLogic Server is installed.
All the nodes of the cluster have identical files sbm-cluster.xml, sbm-properties.xml, and any
application server specific resource files like sbm-resources-weblogic.xml.
Open the file SBM_HOME\cluster\conf\sbm-properties.xml and provide the value for the location
attribute of <property-files> element to point to the location of the conf directory of the SBM_HOME.
Note: If this value is not specified or specified incorrectly, then the property files are located from the CLASSPATH. So, if some other SBM_HOME than the one on which the cluster setup is done, is in the CLASSPATH, then properties of the incorrect SBM installation will be updated.
To modify the SBM configuration files for the cluster configuration, run SBM_HOME\cluster\bin\modifyprop.bat on each node in the cluster to update the properties in the SBM conf files for the cluster. It internally calls SBM_HOME\bin\updateSBM.bat to update the SBM property files in the jar and ear files. This will back up the affected configuration files under SBM_HOME\conf directory by appending .orig extensions to the respective conf files.
Clustering Guide 58 Setting Up the Cluster
In case the cluster configuration file is other than sbm-cluster.xml, you need to run the modifyprop.bat |sh file with the option -f with the configuration file name. modifyprop.bat [-f clusterFileName] where clusterFileName is the name of the cluster configuration file to use. If no file is specified, then sbm-cluster.xml is used
Shut down the Admin server if it is running. Remove the temporary files as mentioned in Before Clustering on page 53.
Clustering Guide
59
Clustering Guide
60
7. To start the BizPulse server, go to SBM_HOME\bin. a. If the EJB Cluster is set up, then start with the command startBizPulseServer -nodes <nodes in cluster> b. If the EJB Cluster is not set up, and there is a single EJB server, then start with the command startBizPulseServer For more information on BizPulse server commands, refer to BizPulse Server Commands on page 78.
Note: BizLogic and BizPulse servers can also be started using BPM Portal.
Clustering Guide
61
where <server name> is the EJB node name, and <admin url> is the URL of the AdminServer specified in the cluster configuration. For example, stopEjbClusterInstance sys97 192.168.0.28:18791 7. In case of a single EJB server, stop the server with the following command: stopEjbServer 8. Stop the Admin server. stopAdminServer
Note: BizLogic and BizPulse servers can also be stopped using BPM Portal.
Performance Tuning
To improve the performance, you can carry out the following tuning: Set the value of PageCheckSeconds and <servlet-reload-check-secs> to -1 in the weblogic.xml file (This procedure makes sure that WebLogic does not check for JSP changes which takes considerable time on Solaris operating system). The weblogic.xml file is located under bea_home\user_projects\domains\sbm_domain\applications\sbm\WEB-INF
Clustering Guide
62
The out-of-the-box installation comes with two versions of WebLogic.xml; namely, weblogic.xml.dev and weblogic.xml.prod, under bea_home\user_projects\domain\sbm_domain\applications\sbm\WEB-INF. The weblogic.xml.prod file already has the PageCheckSeconds and <servlet-reload-check-secs> values set to -1. Rename the file to weblogic.xml and restart the Portal server instance. The original copy of weblogic.xml is the same as weblogic.xml.dev and is tuned for the development environment. You can also check the parameter values through the admin console. 1. Login to Admin console and on the left pane go to deployments -> Web Application Modules -> sbm. 2. In the right pane, click on descriptors -> weblogic.xml
Clustering Guide
63
Performance Tuning
Chapter 7
This chapter describes the details of WebSphere application server required for manual editing of the file sbm-cluster.xml. For basic reference information on manual cluster configuration, refer to the chapter Editing the Cluster Configuration File on page 32. If you plan to use the Cluster Tool instead, refer to the chapter Using the Cluster Configuration Tool on page 18. You can also use the file SBM_HOME\cluster\docs\readme_websphere.txt for quick reference. Clustering in SBM uses the following terminology:
SBM_HOME: Refers to the installation directory of SBM. ADMINSERVER_HOME: The location where Deployment Manager is installed. This is the
physical path where the WASND 6.0.2.5 is installed. For example, D:\WebSphere\DeploymentManager.
CONNTYPE: This determines what connector WebSphere uses. The valid values are RMI and
SOAP. The default is SOAP.
PORT: This is the port used when for connecting to wsadmin. The default SOAP port for a cell
manager is 8879. This may change based on the value provided for the parameter "-startingPort" during creation of the Deployment Manager profile. Currently the value given for "-startingPort" is 4000. This may change if you have both WebSphere base and WebSphere ND installed in the same machine providing the co-exist option. Refer to SOAP_CONNECTOR_ADDRESS::port parameter value in serverindex.xml at WebSphere_Install_root\AppServer\profiles\dmgr\config\cells\sys90Cell\nodes\sys90Manager.
PROFILE: The profile defines the run-time environment and includes all of the files that the server
processes in the run-time environment can change. A profile in WebSphere is similar to a domain in WebLogic.
64
WEBSPHERE BASE: Websphere Application Server Base (WAS Base) is used in the case of
standalone applications. Therefore, for standalone SBM to be installed, WAS Base needs to be installed. This installation creates a default profile with a server "server1". But for installing SBM, we need to create a new profile on which the SBM applications can be deployed.
WEBSPHERE ND: Websphere application server Network Deployment (WAS ND) is used in the
case of WebSphere clusters. If an SBM cluster has to be setup, then WAS ND is needs to be installed. In the previous version (WAS ND 5.1), installing WebSphere ND used to create a Deployment Manager. From version WAS ND 6.0 onwards, you need to create a deployment manager profile, once Websphere ND is installed. This version of Websphere ND also supports creation of WAS Base type of profiles to be created on it. For additional information, and understanding WebSphere terminology, refer WebSphere documentation.
Note: When you install a BizLogic/BizSolo application in a clustered environment, it creates the application files on any one node in the cluster. You need to copy them from the node where it got installed, to all the other nodes in the cluster. You can avoid copying the files manually by using the BPM Studio. Refer to Using BPM Studio in Clustered Environment on page 16 for more information.
Before Clustering
This example refers to a cluster setup of three machines, sys76, sys114, and sys90. The following figure illustrates the architecture of the cluster.
Figure 18 WebSphere - Three-node Cluster
It is also possible to set up cluster with only two nodes. Refer toWorking with Two-Node Clusters on page 74.
Clustering Guide
65
Before Clustering
Clustering Guide
66
Before Clustering
Using Ports
While specifying the port numbers in the cluster configuration file, note the following points:
The entire EJB cluster node should have the same <jndi-port> value. The <jndi-port> number of
the jmsserver should be the same as the <jndi-port> of the EJB cluster node, which is hosting the jmsserver. For example, if the EJB cluster is described as below: <cluster name="sbmejbCluster" type="ejb" create="true"> <nodes> <node name="sys114sbm65node"> <address>sys114.tdiinc.com</address> <jndi-port>9811</jndi-port> <sbm-home>D:/SBM65_WS60</sbm-home> </node> <node name="sys76sbm65node"> <address>sys76.tdiinc.com</address> <jndi-port>9811</jndi-port> <sbm-home>D:/SBM65_WS60</sbm-home> </node> </nodes> <persistence> . . </persistence> </cluster> then, the jmsserver should have the following values: <jmsserver name="sys76sbm65node"> <address>sys76.tdiinc.com</address> <jndi-port>9811</jndi-port> </jmsserver> Note that the address and the <jndi-port> number of the jmsserver are the same as the address and <jndi-port> of the EJB cluster node, which host the jmsserver. Also all the EJB cluster nodes have the same <jndi-port> number.
NOTE: The node name should be same as the one you specified while installing SBM. You can find this node name from the sbm.websphere.node.name parameter in the SBM_HOME\conf\sbmjndi.properties file.
The tag <http-port> can also be specified after <jndi-port> tag for EJB server or EJB cluster to access the HTTP port (if available) on this EJB node. This tag is optional, and not valid for JMS server.
All the Web cluster nodes should have the same <http-port> value. The <http-port> of cluster
nodes of the Web cluster should be same as the <http-port> of the proxyserver. For example, if the proxyserver is described as below: <proxyserver name="sys90" type="websphere"> <address>10.1.32.128</address> <http-port>9448</http-port> </proxyserver>
Clustering Guide
67
Before Clustering
then, the Web cluster should be as shown below: <cluster name="sbmwebCluster" type="web" create="true"> <nodes> <node name="sys114sbm65node"> <address>10.1.32.173</address> <http-port>9448</http-port> <sbm-home>D:/SBM_WS60</sbm-home> </node> <node name="sys76sbm65node"> <address>10.1.32.114</address> <http-port>9448</http-port> <sbm-home>D:/SBM65_WS60</sbm-home> </node> </nodes> <persistence> . . </persistence> </cluster> Note that the <http-port> for all the Web cluster nodes is the same as the proxyserver <http-port>. All the Web cluster nodes have the same <http-port> number.
NOTE: The node name should be same as the one you specified while installing SBM. You can find this node name from the sbm.websphere.node.name parameter in the SBM_HOME\conf\sbmjndi.properties file.
The tag <jndi-port> can be specified for portal server nodes for query service to access the JNDI resources (JDBC, etc.) on the portal server nodes. The <jndi-port> value will be different than the <http-port> in case of a standalone portal server. In case of a cluster, the JNDI tree for all nodes is present at the cell level and hence the <jndi-port> value will be the same as the EJB <jndi-port> value.
Configuring Clusters
Configuring a cluster in WebSphere involves federating the installations with WebSphere ND, setting up EJB and Portal clusters, and configuring Portal load balancing.
Clustering Guide
68
Configuring Clusters
3. Edit SBM_HOME\cluster\bin\setenv.bat|sh on all the nodes to verify WebSphere parameters. Ensure the location specified for the ADMINSERVER_HOME does not contain backward slash (\). 4. Create and initialize the SBM tables in the database. In SBM_HOME\bin on sys76 and sys114, execute: setupSBM.bat -c all -u ebms -p ebms -yes
Note: After setting up the cluster, restart DMgr.
Execute setupcluster.bat|sh -cmd in SBM_HOME\cluster\bin on one of the cluster nodes. Configuring Portal Load Balancing
Configure IBM HTTP Server to perform SBM Portals load-balancing. 1. Edit the configurewebserver1.bat located under the bin directory of the Plugins installation directory. In this file, add the ND installation path before the wsadmin script and configureWebserverDefinition.jacl. For example, wsadmin.bat -f configureWebserverDefinition.jacl webserver1 IHS "C:\\IBM\\IBMHTTPServer" "C:\\IBM\\IBMHTTPServer\\conf\\httpd.conf" 9448 MAP_ALL "C:\\IBM\\WebSphere\\Plugins" unmanaged aruba.savvion.com aruba.savvion.com windows modify it to: C:\\IBM\\WebSphere\\AppServer\\bin\\wsadmin.bat -f C:\\IBM\\WebSphere\\AppServer\\bin\\configureWebserverDefinition.jacl webserver1 IHS
Clustering Guide
69
"C:\\IBM\\IBMHTTPServer" "C:\\IBM\\IBMHTTPServer\\conf\\httpd.conf" 9448 MAP_ALL "C:\\IBM\\WebSphere\\Plugins" unmanaged aruba.savvion.com aruba.savvion.com windows 2. Log in into the WebSphere Admin Console. In the left panel, go to Servers > Web Server, and then in the right panel, select webserver1 and click Generate Plugin-in. After confirmation click Propogate Plugin-in. 3. Open the file sbm-cluster.xml from the SBM_HOME\cluster\conf directory and check the value for <http-port> under the tag <proxyserver>. Copy this value to the value for the property Port in httpd.conf in the IBM Http Server installation location. For more information, refer to WebSphere documentation - "Using WebSphere Application Server V5 for load balancing and failover".
Clustering Guide
70
3. Edit the SBM_HOME\conf\webservices.conf file and update the following properties: uddi.0.inquiryURL=http://jaguar.savvion.com:9448/sbm/uddi uddi.0.publishURL=http://jaguar.savvion.com:9448/sbm/uddi Both the URLs should point to the individual portal server URL. In the example above, http:// jaguar.savvion.com:9448 is the portal server URL of the individual node on jaguar. 4. Run SBM_HOME\cluster\bin\modifyprop.bat on each node in the cluster to update the properties in the SBM conf files for the cluster. This will back up the affected configuration files under SBM_HOME\conf directory by appending .orig extensions to the respective conf files. In case the cluster configuration file is other than sbm-cluster.xml, you need to run the modifyprop.bat |sh file with the option -f with the configuration file name. modifyprop.bat [-f clusterFileName] where clusterFileName is the name of the cluster configuration file to use. If no file is specified, then sbm-cluster.xml is used 5. Run SBM_HOME\bin\setupSbmPortalServer.bat on both sys76 and sys114.
Clustering Guide
71
2. Execute startCluster.bat from SBM_HOME\cluster\bin. The wsadmin tool exits after ensuring the Start command to individual servers, so make sure that all the servers are started by checking the logs before proceeding (the portal server may take several minutes to start). Alternatively, you can individually start the servers as follows: a. On the sys76 node, (APPSERVER_HOME\AppServer\profiles\<profileName>\bin):
2. If the EJB Cluster is not set up, and there is a single EJB server, then start with the command startBizPulseServer For more information on BizPulse server commands, refer to BizPulse Server Commands on page 78.
Note: BizLogic and BizPulse servers can also be started using BPM Portal.
3. Remove EJB and Portal Cluster by executing removeSBMClusterFromWS.bat from SBM_HOME\cluster\bin. 4. Convert the setup to stand-alone SBM installations by removing the nodes that were federated to the Deployment Manager:
Note that the command must not be issued on two nodes at the same time. Execute removeNode on computer sys76 first, wait for the command to complete, and then execute it on computer sys114. After the command completes, you should get a message that the node has been successfully removed. 5. Rename the backed up configuration files from the SBM_HOME\conf directory by removing the .orig extensions. 6. Edit the SBM_HOME\conf\sbm.conf and SBM_HOME\conf\sbmjndi.properties files and set the sbm.webappdir parameter in the sbm.conf and sbm.websphere.cell.name parameter in the sbmjndi.properties file to their original values which you had noted down in the step 1 of the Modifying SBM Properties on page 70.
Clustering Guide 73 Restoring Original Configuration
7. Remove the messaging engine datastore created inside the folder APPSERVER_HOME\AppServer\profiles\<profileName>\databases\com.ibm.ws.sib. The messaging engine folder name follows the following convention: <nodeName>.<clusterMemberName>-<busName>. This has to be done in all the cluster member machines.
You can compare it with the architecture of the three-node cluster illustrated in Figure 18 on page 65. 1. Install WebSphere 6.0.2.5 Standard Edition on sys76 and sys90. 2. Install WebSphere 6.0.2.5 Network Deployment on sys90. During the installation, when a screen prompts "Modify ports for coexistence", select the check box. This is a very important step and if not done properly, the cluster may not work. 3. Federate both the nodes (sys90 and sys76) to the Deployment Manager (sys90). 4. Install SBM on sys90 and sys76. 5. If you need to federate the nodes after installing SBM, the parameter SBM.WebAppDir in sbm.conf needs to be manually changed on each node, using the following steps: a. Find out the node name where Deployment Manager is installed, sys90 in this case.
Clustering Guide
74
b. Change the value of SBM.WebAppDir from individual machine name to the Deployment Manager machine name. For example, in sys76, from D:\WebSphere\AppServer\installedApps\sys76\sbm.war.ear\sbm.war to D:\WebSphere\AppServer\installedApps\sys90Network\sbm.war.ear\sbm.war. c. Make sure that the change is carried out on all the nodes, for example, on both sys76 and sys114 in this case.
Note: When WebSphere 6.0.2.5 is installed on sys90, make sure that IBM HttpServer is installed as a Windows service. Refer to the IBM HttpServer Installation document for more details.
Clustering Guide
75
Appendix A
This appendix describes the BizLogic commands you need to use for all the supported application servers. You can also start BizLogic Server from the BPM Portal Administration tab. Go to System > Status, select BizLogic Server and click Start.
76
Use the following commands to start one or several of all the servers. Assume a cluster with three nodes. One node in the cluster is shut down or crashes. After the node is restarted, use the following command to start only the BizLogic server on that node: addBizLogicNodes u ebms -p ebms -nc 1, where 1 indicates the number of nodes to restart. If you need to restart two nodes, use: addBizLogicNodes u ebms -p ebms -nc 2. If a node has crashed or a new node is added, invoke the following command for Archiver or EventPublisher: startArchiver.bat|sh startEventPublisher.bat|sh If you see a warning message indicating that the Archiver or EventPublisher is already started, you can disregard it.
Clustering Guide
77
Appendix B
This section describes the BizPulse commands you will need to use for all the supported application servers.You can also start BizPulse Server from the BPM Portal Administration tab. Go to System > Status, select BizPulse Server and click Start.
78
If you want to stop BizPulse monitoring without stopping the BizPulse server, you can stop BizPulse monitor by running: Windows: stopBizPulseMonitor.bat -nodes <nodes in cluster> UNIX: stopBizPulseMonitor.sh -nodes <nodes in cluster>
Clustering Guide
79
Glossary
This is a listing of common SBM terminology. For a full version of terms used in SBM documentation, refer to the Terminology Guide in the SBM_HOME\docs folder.
A|B|C|D|E|G|H|I|K|M|P|R|S|T|U|W
A
ACL Manager
Go to Top
In Savvion BusinessManager (SBM), Access Control List Manager provides a finer, more precise control over user access rights for resources and actions.
Activity Workstep
In SBM, the basic unit of work; must be performed by one or more human performers (valid individual user, multiple users or user group).
Adapter
A Java class that integrates remote, third party classes and actions with SBM. An adapter can automate certain functions and tasks performed by a remote server or other external systems.
Administration
A module in BPM Portal enabling the administrator to perform such tasks as installing/uninstalling applications, modifying configuration parameters controlling SBM operations, and manage SBM users, groups and access control.The Administration module is visible only to SBM users who have permissions to access it.
Application
In SBM, an application is an installed, executable business process that automates a business flow.
80
B
Balanced Scorecard
Go to Top
A management application in the Management module that measures performance by analyzing how an organizations business activities help it achieve its strategic goals. The Balanced Scorecard provides an analysis from a range of perspectives.
BAM
Business Activity Management combines Business Process Management with strategic and analytical information on specific business performance indicators, providing real-time status information and identifying critical events to assist senior management in making informed business decisions.
BizLogic
An SBM component that provides a flexible, lightweight, scalable workflow process engine for intranets, extranets, and the Internet.
BizPulse
An SBM component that provides an open event-driven rule engine to formulate and enforce policies in business applications.
BizSolo
An SBM component that enables users to develop customizable, sophisticated presentation flows for business processes, install them as Web applications, and execute them on their Web browsers.
BizStore
An open repository of rich e-business information collected from different SBM modules.
BPEL
BPEL (Business Process Execution Language) for Web services is an XML-based language designed to enable task-sharing for a distributed computing or grid computing environment - including across multiple organizations using a combination of Web services.
BPMN
BPMN (Business Process Modelling Notation) provides businesses with the capability of defining and understanding their internal and external business procedures through a Business Process Diagram giving organizations the ability to communicate these procedures in a standard manner.
BPM Portal
An SBM component that offers users, managers, administrators and developers a unified, customizable portal for single sign-on access to all SBM functionalities to which they are granted permission.
BPM Studio
An SBM component that is an Integrated Development Environment for SBM and enables SBM users to develop and deploy an SBM application without leaving the development environment.
Business Calendar
An SBM feature that accurately calculates the Due Date of tasks, and provides support for multiple business calendars across different time zones.
81
Business Flow
The logical sequence of process activities, related to one another by a triggering activity, to achieve an outcome. It represents a business process that begins with a commitment and ends with the termination of that commitment. In SBM, business flow includes workflow (the flow of all human-performed activities), integration flow (the flow of activities performed by systems) and presentation flow (from a users viewpoint, the flow of data from one Web page to the next).
Business Logic
The control flow and information flow among worksteps that define a business process.
Business Object
A representation of an activity in the business domain, including its name, definition, attributes, behavior, relationships and constraints.
Business Process
A process involving multiple worksteps in the form of operations, interactions and notifications performed by a user, group of users, an external adapter, or a script. Examples of internal business processes are purchasing, expense reporting, help desk, and time card. External processes may include order processing, sales, and customer service.
Business Rule
A combination of elements, including validation edits, logon verifications, database lookups, policies and transformations, that represent an enterprises way of doing business.
C
Cluster
Go to Top
In SBM, an integrated set of multiple, electronically connected but physically separated servers. This loosely coupled group of nodes collaborate to provide shared access to the services that each node hosts. To other systems, the cluster appears as a single system.
Control Flow
The sequences of worksteps and workstep conditions, as defined in a process template in BPM Studio or Process Modeler.
D
Dashboard
Go to Top
An SBM feature that provides a graphic overview of the status of several business processes on a single Web page, enabling users to monitor the progress of each process. Users can view business processes across all applications or for a selected application.
Dataslot
A data placeholder that persists through the entire process and defines the information flow of the business process. Dataslots are associated with processes, where they can add information into (Input type) or out of (Output type) worksteps, and appear as editable or read-only fields on a users interface.
E
Expression Editor
Go to Top
An SBM tool that enables users to define complex conditional expressions within a Decision workstep to support their business requirements.
82
G
Group
Go to Top
In SBM, an entity that has as members valid users or other groups who perform related work and have authorized access to specific components.
H
Heatmap
Go to Top
An SBM feature that provides a convenient, graphical tool for managers to visually locate the bottlenecks in the process execution. It helps managers to get an overview of the status of the currently active instances, identify suspended instances, and analyze the history of the completed instances.
Home
A module in BPM Portal through which users interact with SBM. Using the Home module, users complete entries to various tasks and applications, update profile, set preferences, and link to the support infrastructure required to achieve these tasks. The Home module is the primary interface for SBM application users.
I
Infopad
Go to Top
In SBM, a data structure used to capture business metrics, typically displayed as a table with one or two dimensions.
Instance
An individual object within a specific class. In SBM, a self-contained unit that is created each time you use a process template to run an SBM application.
K
KPI
Go to Top
Key Performance Indicator, used in the Balanced Scorecard system, that provides the data translating enterprise goals into a set of measurable objectives.
M
Management
Go to Top
A module in BPM Portal enabling the managers to query, report, and control processes and resources for users of SBM. The Management module is visible only to SBM users who have permissions to access it.
Managed Adapter
In SBM, a Managed Adapter is an implementation of an adapter interface that facilitates data exchange between SBM processes and external applications.
Migration
The process of moving from the use of one operating environment to another operating environment that is typically seen as improvement. Migration can involve moving to new hardware, new software, or both. It may involve a new application, another type of database, or a redesigned network. Migration is also used to refer simply to the process of moving data from one storage device to another. SBM supports data migration as well as application migration.
P
PAM
Go to Top
Process Asset Management recognizes that a companys processes are a unique asset and provides a structure that organizes, stores, and secures process data, enabling users to retrieve information on any of the companys processes.
83
Performer
An entity that executes a workstep. Depending on the workstep type, the performer can be a human user, a group of users, an adapter or other external performer, or a script.
Presentation Flow
The flow of information and user input from one interface to the next. Typically related to a single Activity workstep in the process and generated in a BizSolo environment.
Process Engine
Orchestrates the execution of business processes and also coordinates conversations among process engines based on public processes, which forms the backbone of global business collaboration.
Process Modeler
A stand-alone component that enables users to design templates for basic business processes and store or retrieve them from the Savvion central process repository, and to design and run simulations of processes.
Process Refresh
An SBM feature for replacing the installed process without versioning, facilitating the running process instances to refresh and seamlessly adapt to the new workflow.
Process Repository
A central place (can contain several databases) in which a group of processes is stored and maintained in an organized way, and can be accessed to retrieve information.
Process Template
In SBM, a model of business flow that includes worksteps, connectors and dataslots. After users deploy and install it as an application in the SBM directory structure, they can use the application to create process instances.
R
Rollback
Go to Top
In SBM, a feature that restarts the workflow from a workstep previously selected as the rollback point in the process, performed automatically in the event of a failure.
Role
The actions and activities assigned to a valid SBM user who is a member of a group. In SBM, only members of a group can be assigned a role. A role indicates the relationships of the user in a group context.
Rule Wizard
An interactive utility that enables SBM users to quickly develop rules that can be applied to a business process.
S
SBM Application
Go to Top
An application is an implementation of a business process. It can contain one or more process templates, performers, adapters, customized forms or rules. An application can be deployed, installed and run on BizLogic servers. In SBM, an application is an installed, executable business process that automates a workflow.
Swim Lanes
Used in workflow diagrams to organize complex processes across functional boundaries. For example, seen as horizontal lines on a process map, swim lanes can be used to place individual task steps into different categories that depend on task ownership.
84
T
Task
Go to Top
In SBM, a performer is assigned one or more workitems that the performer sees as tasks. There are two types of tasks: Assigned, which are assigned specifically to you; and Available, which are available to be performed by you or other members of your user group.
U
User
In SBM, a valid human performer with authorized access to specific modules.
Go to Top
W
Workflow
Go to Top
The logical sequence of activities performed by human performers, typically in a BPM Studio or Process Modeler environment. Workflow includes the tasks, procedural steps, organizations or people involved, required input and output information, and tools needed for each activity in a business process.
85
Index
A
applications deploying on a cluster 16 Archiver 77
D
deploying applications on a cluster 16 Designing Server Clusters 11
B
batch script 76, 77 BizPulse Clustering Failover Configuration 48, 59, 71 BPM Portal Clustering 15 BPM Studio Clustering 16
E
EJB Cluster 56 EJB server 56 Environment Variables Setting up 56 EventPublisher 77 Excel Adapter, issues in clustering 17
C
Cluster Tool 18 Cluster Panel 20 Configuration Panel 20 Server Panel 21 Setting up 21 Starting 19 Clustering 10 Adding New and Recovered Nodes 30, 41 Designing 11 for Failover 14 for Load Balancing 14 in SBM 12, 14 Terminology 10 Clustering Framework 18, 32 Cluster Information 21, 37 Cluster Setup in Stages 40
F
Failover BizPulse on Pramati 48 BizPulse on WebLogic 59 BizPulse on WebSphere 71
I
installing SBM as Windows Service 13 iWay Adapters, issues in clustering 17
86
J
JMS server 56
L
License Agreement 2
M
managed adapters in clustering 16
P
Product Support Team 9
R
Restoring Original Configuration Pramati 47 WebLogic 59 WebSphere 73
S
SBM conventions in documentation 9 documentation list 8 user types 7 Setting up EJB Cluster Pramati 46 WebSphere 68 Setting up JMS Cluster WebLogic 56 Setting up Web Cluster Pramati 46 WebLogic 57 WebSphere 68
U
URL Document Server 28, 40 Using Ports Pramati 44 WebLogic 53 WebSphere 67
W
WebSphere Three-node Cluster 65 Two-node Cluster 74 Windows Service, installing SBM as 13
Clustering Guide
87