Additional System Requirements for a Distributed BPM System
A distributed ActiveMatrix BPM system has additional architectural requirements.
An example of a distributed BPM system is shown in Example configuration for a distributed ActiveMatrix BPM system. The requirements are shown in the following table.
Requirement | Description |
---|---|
Load Balancer | Manages HTTP requests from ActiveMatrix BPM clients across the machines hosting BPM nodes. |
Shared File System | Holds shared configuration data used by ActiveMatrix BPM and/or ActiveMatrix Administrator. |
LDAP Server | Must be external. Do not use the internal LDAP server supplied with ActiveMatrix BPM. |
For a high availability configuration, they must also each be configured to provide high availability and fault tolerance.
Load Balancer
The load balancer must be configured as follows:
- The following ports must be configured for load balancing across the relevant machines:
- ActiveMatrix Administrator external HTTP port (default 8120) - used for communication between ActiveMatrix Administrator and external web/CLI clients and Business Studio.
- ActiveMatrix BPM HTTP port (default 8080) - used for communication between ActiveMatrix BPM and external clients (Openspace, Workspace and custom clients).
- Sticky connections should be used so that client session context is maintained. (If the server hosting this session dies the client will be logged out and need to log back in again.)
Shared File System
The shared file system must:
- be available to each machine that is part of the BPM system.
- be mounted in an identical location on each machine. For example, if you create a network drive or folder on machine 1 and map it to x:\shared on the primary machine , the shared drive or folder must be mapped to x:\shared on each remote machine as well.
- contain the folders shown in the following table (where SHARED_DIR is the root folder of the shared file system).
Folder | Notes |
---|---|
SHARED_DIR\admin\enterpriseName\shared | Shared staging area folder used by ActiveMatrix Administrator.
Note: This folder is only required for a high availability configuration, where the Administrator server is replicated on two machines.
Use this folder when configuring the ActiveMatrix Administrator on the primary machine - see Primary Machine: Moving the ActiveMatrix Administrator Server Staging Area to the Shared File System. |
SHARED_DIR\bpm_config | Shared configuration folder in which ActiveMatrix BPM stores BPM properties files. (bpm_config is the suggested name for this folder, but you can use any name that you wish.)
Use this location when configuring ActiveMatrix BPM on:
|
SHARED_DIR\keystores | Shared authentication folder in which ActiveMatrix BPM stores keystores (or trust stores) used for SSL and/or WSS authentication. This folder is created automatically if you select
Override BPM Configuration Folder in the
ActiveMatrix BPM: Configuration screen of the
Create TIBCO ActiveMatrix BPM Server Wizard.
To ensure that authentication operates correctly, you must:
|
Server Time Synchronization
The primary and remote machines must have their time/date synchronized with each other. If server times are offset (even by a few seconds), the ActiveMatrix BPM node log file:
CONFIG_HOME\tibcohost\Admin-enterpriseName-adminServerName\data_3.2.x\nodes\nodeName\logs\BPM.log
may show significant numbers of errors like the following:
05 Jun 2011 14:12:02,465 [QuartzScheduler_PVM:Quartz Scheduler-LINAramis1307296618457_ClusterManager] [WARN ] com.tibco.pvm.persist.quartz.jdbcjobstore.IpmQuartzJobStoreTX - This scheduler instance (LINAramis1307296618457) is still active but was recovered by another instance in the cluster. This may cause inconsistent behavior 25 Jan 2012 01:33:48,698 [QuartzScheduler_PVM:Quartz Scheduler-mjkmssqlbe1327088686491_ClusterManager] [INFO ] com.tibco.pvm.persist.quartz.jdbcjobstore.IpmQuartzJobStoreTX - ClusterManager: Scanning for instance "threemjkv121327089184017"'s failed in-progress jobs. 25 Jan 2012 01:33:56,200 [QuartzScheduler_PVM:Quartz Scheduler-mjkmssqlbe1327088686491_ClusterManager] [INFO ] com.tibco.pvm.persist.quartz.jdbcjobstore.IpmQuartzJobStoreTX - ClusterManager: detected 1 failed or restarted instances. 30 Jan 2012 09:43:38,067 [QuartzScheduler_PVM:Quartz Scheduler-twomjkv121327596401015_ClusterManager] [WARN ] com.tibco.pvm.persist.quartz.jdbcjobstore.IpmQuartzJobStoreTX - This scheduler instance (twomjkv121327596401015) is still active but was recovered by another instance in the cluster. This may cause inconsistent behavior.