Kbase P105542: How to install Fathom Management on Cluster ?
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
12/12/2006 |
|
Status: Unverified
GOAL:
How to install Fathom Management on Cluster ?
GOAL:
How to make Fathom Management cluster-aware ?
FACT(s) (Environment):
Progress 9.1E
OpenEdge 10.0X
Fathom Management 3.X
Windows 2003
FIX:
This solution assumes that your OS Cluster software is working properly and also Progress 9.1E/OpenEdge 10.X has been properly configured to work on a Cluster (See solution P102918, How to make AdminServer cluster-aware ? for more details)
Steps taken on the active node:
1) stop the Progress AdminServer + databases
2) install Fathom Management 3 to a local folder
3) install latest SP for Fathom Management (not mandatory, but highly recommended)
4) on the shared (Quorum) drive (Q for future references), create subdirectory Fathom with following subfolders: config, db, work
5) start the AdminServer on the active node
6) access the Fathom Console and start configuring Fathom (Fathom db, email, ....)
7) when prompted for location of Fathom DB, select Q:\Fathom\db\fathom
8) finish the configuration and after it's done, stop the Progress AdminServer
9) edit $FATHOM/fathom.init.params and modify following lines (change the location from local drive to Q: drive)
SET fathomConfigDir=Q:\Fathom\config
SET fathomLogsDir=Q:\fathom\work\logs
SET fathomWorkDir=Q:\fathom\work
SET fathomReportDir=Q:\fathom\work\reports
java.security.auth.login.config=q:\Fathom\config\loginmodule.config
and save the file
10) move the content of $FATHOM\config directory to Q:\Fathom\config
11) move the content of $FATHOM\wrk directory to Q:\Fathom\work
eg: the directories will still be present on local drive, but their content will be moved to shared drive
Steps taken on the standby node(s):
0) since it's a standby node, no AdminServer, no db is running
1) install Fathom Management 3 to a local folder
2) install latest SP for Fathom Management (not mandatory, but highly recommended)
3) deploy content of $FATHOM/etc directory from primary node to $FATHOM/etc on standby node(s) - override the current content
4) deploy $FATHOM/fathom.init.params from primary node to $FATHOM on standby nodes(s) - override the current one
5) delete content of $FATHOM/config directory on standby node(s) - only the content, the directory remains
6) copy following files from $DLC/properties directory from primary node to $DLC/properties on standby node(s) - override current ones: ubroker.properties, conmgr.properties, AdminServerPlugins.properties, JavaTools.properties, management.properties
Now the AdminServer/Fathom Management are configured on Cluster - the Fathom Trend database, resource database, working directory and configuration files are shared.
Things to bear in mind:
1) anytime you will add a new user to a Fathom Console or you change a password for current one, it is necessary to deploy content of $FATHOM/etc directory across the Cluster (to all standby nodes)
2) during the failover, Q:\Fathom\config\fathom.odx directory does not get deleted which prevents Fathom to start - this is due to fact, that shared device is moved to another node first and AdminServer does not have a "time" to properly close it. It is recommended to create a script/batch which will delete the content together with fail-over.