Jon Aimone wrote: > Hi, > > I've been reading through the AI documentation, and it's not clear to me > how one would implement the "old jumpstart" feature of derived profiles. > I'm certain I must have missed it...
No you didn't miss it, its not there.. > > Currently we have a large lab populated with a wide variety of systems > -- no two alike. We use derived profiles and a "smart" begin script to > generate the profile based on the hardware present... choosing the boot > disk (because it's the one on the particular type of HBA desired, versus > simply choosing c0t0d0 or something based on size), selecting multiple > NICs to configure, even including additional software packages based on > the HW configuration. > > Needless to say our begin script is rather extensive, our rules file has > twelve lines in it and never changes, and our finish script actually > pre-configures some of the software installed via the derived profile. > > How do I accomplish the same with the AI? client-side processing to derive a manifest is not supported, and not yet on the map for AI. The current design requires custom manifests to be pre-built and stored on the server with the evolving criteria specification as the selection mechanism. I can understand how this may not meet your requirements though, and would like to see where we can make enhancements to bridge this. Can you give more detail on the type of variables you base your customization choices on? Perhaps just attach your begin script? thanks, -ethan
