If the disks are seen using just a couple of SATA ports on the node, maybe you need to check the BIOS settings (assuming there's nothing wrong with the cabling!)
I have gone through the BIOS setup of one of the nodes. I have 12 disks so each node has 4 SATA disks connected. The images show how my BIOS is setup. The main thing to check is the Southbridge Config (in "Chipset" folder) and that the "SATA IDE Combined mode" is set to DISABLED. The comments at the right of this item are the key. If you want to have upto 6 SATA ports (that's the max onboard anyways), it should be set to DISABLED.
Tried to check the user guides from DELL for similar boards (not the same...) but even these docs did not have the pin out for the power connectors. Maybe a h/w manual has it. The problem with these boxes is that they are custom made and we can't find decent (aka detailed) documentation. Neither Dell nor TYAN could provide something (I asked, both of them).
Even if you figure it out, you should take into account the cooling of the board. Seriously, my biggest issue (unless you run it in a cool/AC room) is the temp of "SR5650" (it's the main chipset, if not mistaken). The limit is 98 Celsius (!!) and board #2 (bottom right) has temps very close to that limit when I used it in my office (~24 Celsius ambient temp). I had to shut down that board after it started a continuous beep (aka >98 Celsius)Thanks Thanos! Looks like I was editing my post while you replied! Anyway, I have a pair of low profile Swiftech waterblocks on order to cool the processors, but I haven't paid any thought to the chipsets. They seem like they are typical 40mm 2 screw, so while it would add to the BOM, it is more copper pipe I get to runTried to check the user guides from DELL for similar boards (not the same...) but even these docs did not have the pin out for the power connectors. Maybe a h/w manual has it. The problem with these boxes is that they are custom made and we can't find decent (aka detailed) documentation. Neither Dell nor TYAN could provide something (I asked, both of them).
So, I guess you're a pioneer
just FYI. Good luck with the power!
watercooling! hm... good ideaThanks Thanos! Looks like I was editing my post while you replied! Anyway, I have a pair of low profile Swiftech waterblocks on order to cool the processors, but I haven't paid any thought to the chipsets. They seem like they are typical 40mm 2 screw, so while it would add to the BOM, it is more copper pipe I get to run
This is not a design for rooms that are not A/C'd. Yes, the chipset must be those two smaller ones closer to the back of the board. My guess is that the airflow back there isn't enough to cool them if the ambient air temp is higher than 21-22 Celsius (I was OK when I turned on the A/C ). It would be quite interesting to see photos of your finished water-cooling project 
The HD bays are each wired in sets of four to the nodes, and the cables are super easy to reroute if you want to put all 12 to the same node. Each board has 6 SATA ports, so with a card and a riser it'll be easy to centralize the storage. I was debating doing this myself before I decided to embark on this silly build.
That is very similar to the issue I was having on one of the physical nodes when trying to install vSphere 5.5 back in February. I was indeed able to finally get it to work with a link suggested in one of the earlier posts: Dell CS24 ESXi 5.5 Install Stuck "Relocating modules and starting up the kernel..." | RobWillis.info
I actually went through much of this myself back in February (see my posts) and I sympathize with your ordeal. I ended up getting a bare-bones server from the seller on EBay just to use as a parts server. I replaced the problem node with one of the nodes on the bare-bones server back in March and it has worked well for me since. It is very useful having this home lab setup with these servers. I will say I have had to replace 2 hard drives already and I just lost a Baracuda ES2 750GB hard drive today. The first 2 losses were on the same drive bay but the loss today is on a different drive bay. I guess I do expect to loose a few drives occasionally when trying to run a home data center (smile). Having the data center though has greatly helped me to learn a lot about networking, Linux, and VMware. So I guess it has been worth the setbacks that I had when I first started back in February.Status update for whom it may be useful
two of the nodes are setup with ESXi 5.5 update-1. One of them is functional in terms of the compute part as well as the Web access of the IPMI interface. The second node is functional only for the compute part. The IPMI interface is responding only to ipmitool (from a Mac) but not via the Web interface. I have no clue how this can be fixed. Both nodes can see the 4 disks attached to them. No issues there. And I have ESXi installed on a USB (Sandisk Cruzer 8GB) in the onboard USB port. Now, the 3rd node is useless so far. The funny part is that it's IPMI is OK (web and use from a tool like 'ipmitool'). But when it powers on, there's no VGA output and there's continuous 'beep' till I power it off.
I removed the 'button' battery to reset the BIOS to the default settings. Nothing.. IPMI reports "VBAT = 0 Volts" while the other 2 nodes have a value and reported as Normal this is reported as "Lower Critical".
Without documentation I have no clue what it is. I have opened a ticket to TYAN for help on the S8208 board so as to make sense of all this. They replied after a couple of hours (!) saying that I should seek support from DELL since this board was made for DELL. Not very helpful since Product Support from Dell doesn't recognize the Service Tag of this box...
I will open a ticket with Dell. Just for the fun of it. It's a big disappointment that it's so damn hard to get a proper manual for a board. If anyone has a user guide for *this* specific board please share it.
I'm expecting a USB-to-serial cable tomorrow to connect it the COM port of the board. Just in case that I get something done via the old days ways...
By the way, since some guys have expressed the thought to install and experiment with VMware's VSAN on these boxes, wait for 5.5 update 2. It appears that the AHCI driver will be "updated" in Update-2. I could not get it to work in 5.5 update 1. (Kernel times out on trying to use the SATA drives...). Other than that, I think this server is a bargain simply for the amount of DRAM (in virtualization it's the best asset to have
I will report back with news on this front. Or I will purchase an extra barebones server to replace the failing node...