I'm hoping I can destroy and rebuild the array back into a healthy state by swapping around the disks.
Ok, so I have some final results from the last week or so of troubleshooting. I appears that enabling ACLs on my RAID volume in FreeBSD is the source of all the trouble. The ACL read/write errors only appear (in a year's worth of system logs) on the day I enabled and created my ACLs.
[~]# setfacl -d -m u::rwx,g::r-x::o:r-x,mask::rwx Music/
[~]# setfacl -m g:Managers:rwx Music/
[~]# setfacl -m g:MusicManagers:rwx Music/
Just to make absolutely sure this was the case I created a new array from 3x 500GB disks and copied over 50GB+ w/o ACLs applied. No issues at all, everything works great. Then I enabled and applied my ACLs, started another 50GB+ copy operation & tailed the systemlog. Immediately, this shows ACL & superblock errors and, if there are enough read/write operations occurring, this causes the array to collapse and the machine to automatically restart (BIOS reports on restart 'partial array' or 'recovery state').
Not a lot of options here. I have outgrown the primitive *Nix access controls and need the sophistication ACLs provide so, as much as it pains me to do so, I had to retire this server and move the array (4x 2TB RAID5) to my other server (Windows 2008 Server).
For anyone wondering, I was running FreeNAS v0.7.2 (which I absolutely love and hope some day to have the opportunity to leverage again)