Curiously this issue (same OS though) came up again for me this weekend. I would recommend that, instead of simply copying the path code, that one make the INF change and do an actual deployment. That way, one has records of the deployment and is not using a workaround outside of the toolset.
[ QUOTE ]
No, a sun solaris app server package can NOT be deployed to an intel server app server - no "different" server package can be deployed to different servers since the compiled code is different under each platform.
As I was mentioning to John Kozel earlier - if you have multiple app servers on the same OS (and, preferably, on the same OS Version and compiler version) - you can submit a build and deploy a package on a single app server (say a HA machine or a test/dev box), - and then copy the app server code from that box to the production boxes. This makes sense in an environment with many application servers (in Johns' case, he has 6 application servers) - and can be scripted (of course you have to stop services, copy the code then start services again). The script can be run in a way that will minimize downtime to the end users - and, if you use a load-balancing mechanism for your application servers (see my whitepaper for an example) then you can theoretically even deploy code "live" (which, of course, you wouldn't do because you wouldn't want two different versions of the same report being submitted at about the same time on two different app server !)
Alternatively, with only about 2 or 3 application servers, it makes more sense just to submit the package and define each application server.
Its very late - if either you or JK need more info, just ask questions...and I'll try and answer more tomorrow.
[/ QUOTE ]