In previous Solaris releases we relied heavily on client-side extended custom hardware probes and evaluation to provide values for substitution in derived profiles.
This allowed us to create a rather generic templates for the profile using [shell] variables and still be able to customize to the machine's physical characteristics. We actually evaluated each line of the profile as if it were a shell line. -> swap/crashdump space the size of physical memory -> a separate /var/crash file system large enough for several crashdump files -> create hardware RAID volumes large enough to handle the swap/crash space, if necessary and use it as the install target -> choose the target disk(s) for installation based on additional selection criteria. e.g. HBA model, SAN type, storage mode (ST9990V, ST6140, etc.) I'll post more later as I can get it into cogent thought. -- This message posted from opensolaris.org
