Building first render farm

O.k.- so I am in the middle of building my first render farm, and computer for that matter. I was inspired after reading an article on a 32-thread machine using 2nd hand Xeon processors. I was also reading a forum post by another member building a very similar node.

I am done placing and hooking-up all the hardware. I plugged it in last night and hit the power button. The fans came on and things seem to be getting power. I let it just run for about a minute before powering it back down. I am assuming I didn't do anything bad by doing this???

Here's where I am hoping for some feedback:

Tonight I am going to hook-up a monitor, keyboard and mouse and start the software install.

1. I am preparing a flash drive with a bootable version of windows 10

2. plug it into a USB 2.0 hub

3. power up the system and it should go right into the windows 10 install process

4. At some point in the install I will have an opportunity to name the drive letter so I can change it from being "C:" like my desktop computer....

5. Set up Remote Desktop Connection so when I unplug the monitor, keyboard and mouse, I'll be able to log back into the machine remotely

6. Install 3ds max and backburner

Done.

Right?

Sure, the basics are down as you describe them pretty much.

No, there are no issues having your machine running without an OS installed or spending few minutes (or hours) in the BIOS / initial post screen.

Well something isn't working. I fired up the node to let it go through the UEFI Setup Utility and it seems the system is only detecting one of the CPU's.

Can you post your components? Is this a new mainboard or used? Are you already running the latest BIOS?

Did you check the stepping of the CPUs?

You could try them separately to see if the second CPU is ok.

All PSU cables connected?

I think most likely it is a defective board.

MB is new. BIOS is Version p1.8 (seems to be latest). The stepping on the CPU is the same. I actually replaced the first pair I got because I accidentally ordered the SR0KX

I am about to swap the CPU's and see The one in the CPU-BSP1 is working. The CPU-AP1 is not. When I disconnect the power from the BSP1 the machine won't boot at all, just clicks over and over.

Motherboard: ASRock Rack EP2C602-4L/D16 Dual LGA2011/ Intel C602/ DDR3/ SATA3/ V&4GbE/ SSI EEB Server

CPU: (2) Intel Xeon E5-2670 SR0H8

RAM: G.SKILL Ripjaws Z Series 64GB (8 x 8GB) 240-Pin DDR3 SDRAM DDR3 1866 (PC3 14900) Desktop Memory Model F3-1866C10Q2-64GZM

Cooling: Cooler Master Hyper 212 EVO - CPU Cooler with 120mm PWM Fan (RR-212E-20PK-R2)

Power: EVGA SuperNOVA 850 G2 80+ GOLD, 850W ECO Mode

SSD:PNY CS1311 240GB 2.5" SATA III Internal Solid State Drive (SSD) - (SSD7CS1311-240-RB)

SR0H8 = C1

SR0KX = C2

but this shouldn't be the problem here as long as they are both the same.

I think there isn't much you can do. Since you're getting into BIOS, it shouldn't be a RAM problem.

You could try a different PSU. When you swap the CPUs you should check the socket for bent pins.

I think if both CPUs are ok it's the board or the PSU.

O.k.- so first off I mixed up my invoice for the CPU's. I am working with the SR0KX-C2. I returned the SR0H8-C1, but yes you are correct it shouldn't make a difference.

I swapped the CPU chips and the CPU_BSP1 is still the only one being detected. So therefore the chips are fine. The pins on the MB look good too.

I don't have an optical reader for the render node, so I can't use the install disc unless I go and buy one and hook it up. I'd rather not, but is it even worth it? If it's loading the UEFI Setup Utility and detecting the most up-to date BIOS is the Disc going to help me anymore?

FRUSTRATION MANAGEMENT MODE=1

No, the disc is not needed. The latest drivers should be online.

The latest BIOS is 1.80. And i think normally even an older BIOS shouldn't be a problem at all with such "old" CPUs.

Did you try a CMOS reset?

As i said above... board or PSU. So if you have another PSU you could test it. But i think most likely it is the board.

So The reset CMOS didn't do anything to fix it. My last hope is to swap the actual power cables from the power unit and see if the CPU2 power outlet on the EVGA is working, or if the actual cable itself is defective. Was reading on another forum someone was having a similar issue with a Dual MB and after replacing the board TWICE finally traced it back to the power unit's individual output socket itself. Regardless, I've been in contact with the vendor I purchased the MB from and they are in agreement it is most likely the mother board and are being very helpful in working with to exchange it for a new one if the EVGA power unit ends up not being the issue.

UPDATE

So I ended up returning the ASRock EP2C602-4L/D16 motherboard. I was able to receive a full refund after being made to jump through several hoops.

I ended up ordering the ASRock EP2C602 SSI EEB. Now the system won't even POST. I replaced the motherboard battery and did a full CMOS recovery after unplugging, clearing any residual power and set-up the jumper on the ME recovery for 4 hours. Now the ASRock support person is saying the RAM I have hasn't been validated for that board, even though it's the same exact set up as the one described in the article that inspired me to build the render farm node in the first place.

I'm ready to throw the entire thing in the garbage, but I keep reminding myself to maintain patience.

Woot, don't know what to say...empathy is present, but experience with this kind of troubleshooting is not.

Try with just 1x CPU and 1x dimm maybe?

I would also say try with one or two DIMMs first and then add more - i would keep the 2 CPUs installed. I also had the problem with one re-used DDR3 kit that only worked with 1-3 sticks (4x4GB Corsair XMS3 DDR3 1600MHz, which i planned to combine with another set to get 32GB) The system didn't boot with 4 modules. I ended up buying another cheap set of 8x8GB Samsung DDR3 ECC 1333 MHz on ebay like for two other E5-2670 v1 systems and they worked.

So, per the Support reps request I have removed all the RAM except the one in the 'A' slot and removed the second CPU. I then tested each RAM individually and all of them allowed me to boot and get to the pre-POST start-up screen. Although 3 of the sticks gave me a "b4" USB error code. Patience.