lscs and lscs-static

Adam Stoller

Environment[indent]TS/SP/LS/OD 7.2.1 on Linux with WebSphere[/indent]Query[indent]Can someone provide definitions that distinguish "LSCS" from "LSCS-static"?

Rick Poulin

Can someone provide definitions that distinguish "LSCS" from "LSCS-static"? For what is each app responsible?

LSCS is the core app and is responsible for everything you'd consider "dynamic", including executing queries and importing content. When you request a specific document (the syntax that asks for the file content, not the file metadata), lscs does an internal forward to lscs-static. The latter is a "fake" application, in that it's nothing more than a webapp context root pointing to your LSCS-Store, so that the webapp knows where to get the static files. The servlet container is the piece doing the I/O from disk, lscs-static is only configuration information to tell it where that is.

I understand that LSDS takes care of rendering .page files and such.

Nope. This is the biggest grey area in Iwov's marketing spiel. Even when you have LSCS, LSDS completely ignores its existence unless you have specific requirements to use it. In particular, page files are still rendered from the deployed LSDS context root. LSDS will NEVER get static file content from LSCS unless you have Externals or other code that specifically goes to it. LSCS is there for two things: executing rules (and the rest of targeting stuff), and executing content queries. That's it. If you wanted, you could build an External that gets a DCR XML from LSCS, but it's often faster just to get it from your deployed context root. You could also print your URLs to static resources to use LSCS URLs, but there's no value in that when Apache can do a much faster job of serving those up.

Do these apps pre-process content for IDOL or does IDOL reference these apps for resolving queries or what?

These apps are 'dumb' apps, meaning zero business logic. Your PLC does the "pre-processing" of metadata. LSCS simply takes the output from that as an idx file plus payload and sucks it in. There's no additional processing done by LSCS. Also, the OEM IDOL here doesn't actually index your original files. It's not even aware of their existence, because it's only reading the digested IDX which contains nothing more than the metadata you can see as the result of LSCS queries.