CRM - Verity - App server and Process scheduler on separate boxes, how to setup? (Doc ID 618468.1)

Goal

Comprehensive resolution for what all drives to be mapped for Verity to Build Search Collection, when App server and Process Scheduler are installed on separate machines.

Scenario
Customer has two separate physical machines, one for Process Scheduler (ADNTAS44-Machine name) and other for Appserver (ADNTFS99-Machine Name).

SOLUTION:
When Verity Build Collection is schedued to be run on Process Scheduler (ADNTAS44), collections are built on this machine.  At runtime from PIA (goes thru Appserver Machine (ADNTFS99), VQL has to be sent to retrieve information from Collections.

Verity collections are stored as flat files as opposed to Database Tables.  In our Scenario, because Appserver is on a different machine than the Process Scheduler, we need to share the Process Scheduler Machine's drive to the Appserver Machine.

How to Setup/Configure the share drive
1.  Get to ADNTAS44 (Process Scheduler Machine)
2.  Assume the Drive C: on Process Scheduler Machine is where the Collections are built. Within Windows Explorer goto the properties of the drive and make it shareable.
3. Get to ADNTFS99 (App Server Machine)
4  Get to Windows Explorer and create logical Drive  (say H), mapping to Process Scheduler Machine Sharable drive (\\ADNTAS44\C$, for example)
5. After this is done, get to the Appserver config file and add this enrty under [Search Indexes] section (referring to H: drive and collection location). This establishes the link between Appserver and process scheduler machine with regards to collection location at runtime for doing searches.
************************************************************

3) We also need to have a mapping on Data\search directory.
Process Scheduler should be mapped to <PSHOME>\data\search folder on appserver -- so that process scheduler can write the collection on the app server.

So now you will have (with drive name examples):

AppServer
F: Mapped to the Process Scheduler \Data\Search\
V: Mapped to a Shareable Drive to Process Scheduler e root level like c:

Process Scheduler
D: \Data\Search
V: Points to the Shareable Drive on the AppServer

4) Now how to setup the Search Index PATHs in configuration files:
The PIA(Search Collection Run Control) points to drive D in the build process.
The App Server Configuration, psappsrv.cfg, points to App Server Collection Path.
The Process Scheduler Configuration, psprcs.cfg, points to V.
(Please follow the PATH example given at each location, to enter your correct paths)

5) The appserver user and the process scheduler user should have all read and write permissions on each other. The user who has started tuxedo (BEA Proc Mgr service) on the application server machine must have access to this share because no matter what user login you use to start the app server when app server internally tries to access the Mapped drive it will use the login of the user that has started tuxedo.


6) Check if the mkvdk.exe and other verity executables are present on App server box. They should be as they get installed with ptools. And if they are, then remove all these files from process scheduler box.

Once you are done with all the above set up, Build a fresh collection and test Searching for Solutions from Solutions Advisor and Query Test.

Solution

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.