Well we can't tell you the best solution without knowing what exactly the server is doing, however, we can give you a little more general guidance.
In very general terms, certain types of data lend themselves to better performance on certain types of RAID arrays, and I'm only comparing 2 redundant array types, i.e. RAID 1 (or 1+0, 0+1, 10 call it what you want) versus RAID 5.
Binaries, and other sequential r/w files, like O/S files, executables, and logfiles including database logfiles, lend themselves to RAID 1/1+0.
Dynamic or random access files, like database data files and other user data files tend to perform better on RAID 5.
For your DL380, and again we don't know what this server role is, but one example may be to RAID 1 the first to disks giving you about 135GB of RAID1 capacity for O/S, Program Files, and any logfiles including database logfiles. You may or may not split this into multiple NTFS volumes in Disk Mangler (I would, like a 30GB System volume and the rest a separate volume for the rest of your RAID1 stuff). A RAID5 volume accross the remaining 4x disks would yield approx 400GB RAID5 capacity for your database data files, user data, and other dynamic data, and again you may split this into multiple NTFS volumes in Disk Mangler or not.
Your O/S swapfile or paging file really doesn't belong on any redundant volume, but rather than leaving a whole disk out of the mix, many choose the lesser of evils and put the O/S swapfile on a RAID1 volume.
In any redundant RAID, you may only only fail 1 disk at a time. Even if you have a hot spare, if a failure were to occur, the subject array would have to rebuild onto the hot spare before you could afford a second failure.
Hope this is helpful.