CPIT has acquired something of a reputation for being overly fussy and fragile.
Last week I found myself needing to get hold of a FileNet 5.2.1 single server system, and CPIT seemed to fit the bill. Luckily I had all the packages to hand, so within about 3 hours, I had what was at first sight a complete environment.
How wrong I was.
I’d tested the installation out in WorkplaceXT and ACCE, and all was looking great. However, on configuring a repository inside Content Navigator, it all descended into a bit of a mess. Navigator insisted that the object store was not available, and the WebSphere logs gave no hint of any problem – showing successful object store authentication, so, after a few hours I was beginning to wish I’d installed everything manually.
However, after a tip-off (thanks Colin) it became obvious that a critical step had been skipped by CPIT – this is a setting that is present in WebSphere 8.5 that was off by default in earlier versions. Despite overcoming this obstacle, I also found that I couldn’t open any of the Workflow-based Java applets in WorkplaceXT. A bit of googling revealed an article that suggested further changes to WebSphere configuration, and thankfully this fixed the problem.
So now I have a fully operational 5.2.1 system, but I’m left wondering why it had to be so frustrating. Did I miss something in the CPIT documentation? Or, are we just expected to figure out for ourselves that important steps have been missed by CPIT? CPIT is completely scripted, so maybe there is no command-line statement available for making these changes I mentioned above, but even so, the CPIT Installation guide should include these gotchas.