My First DIY NAS Build: Supermicro Technical Support

I was hoping that things with my first DIY NAS build would go smoother. My request for a complete list of the qualified processors for the Supermicro X10SLL-F motherboard used in my first DIY NAS build was clear enough. Milton Cai did me a partial favor in sending along a partial list of those processors. Notably absent from his original list was any mention of the 4th generation Intel Celeron series, as well as all members of the Intel Xeon E3-1200 v3 processors.

I’ve reproduced Milton Cai’s original list, appending the launch date for each product.

Milton Cai’s Original List

G3220 – Q3’13
G3220T – Q3’13
G3420 – Q3’13
G3420T – Q3’13
G3430 – Q3’13
i3-4130 – Q3’13
i3-4130T – Q3’13
i3-4330 – Q3’13
i3-4330T – Q3’13
i3-4340 – Q3’13

Launch dates and additional information on Intel products is always available through Intel’s ARK.

I was surprised to find that the list of qualified processors that I received shows only processors released in Q3’13. Intel continued to release new Haswell processors utilizing socket H3 LGA1150, with support for ECC memory, each quarter (Intel ARK)

I know that there’s a more complete list, so I emailed Supermicro tech support at [email protected] at 2:29PM CST on August 20th:

I’ve called in about my X10SLL-F, updated to BIOS rev 2.0, and issues that I have getting it to POST with Intel Pentium G3258.

I would like to have a most recent CPU compatibility list for this motherboard to help me in selecting an appropriate CPU.

Please send over a full list

I received a response from James Wang at Supermicro at 9:17PM on the same day:

Intel® Xeon® processor E3-1200 v3 and 4th Gen Core i3, Pentium, Celeron processors
Single Socket H3 (LGA 1150)

Regards,
James

This response. Words cannot explain how many things is wrong with this. I read it, sighed with exasperation, and reached for the copy of Swanson’s Rules of Management (or The Unwritten Laws of Engineering) that I keep suspended on the side of my desk. I picked it up and gesticulated wildly to bullet point #5:

Viewgraph rule: When something appears on a viewgraph (an overhead transparency), assume the world knows about it, and deal with it accordingly.

This fucking engineer was either trying to be clever, or he is painfully ignorant: the words he relayed are publicly available and were already considered prior my decision to purchase the Supermicro X10SLL-F. Had he taken a moment to read my message before copy-pastaing this tripe to me, he would been better able to address the situation.

I went to sleep, determined to get to the bottom of this. I woke this morning after a brief respite and shot this back to James Wang.

This is not what I asked for. Can you send me a full list of the qualified processors that your lab has tested on the X10SLL-F? I need this in order to select a CPU that fits my requirements.

The information that you sent me is already available to the public on the product information page: http://www.supermicro.com/products/motherboard/Xeon/C220/X10SLL-F.cfm

I believe that I was clear in stating that I could not get my Pentium G3258 to POST. This is a processor that was released in Q2’14: http://ark.intel.com/products/82723/Intel-Pentium-Processor-G3258-3M-Cache-3_20-GHz

Again – I need the full list of qualified processors for use on the X10SLL-F motherboard. I do not want to purchase another processor only to find out that it does not work.

I sent a similar message to [email protected] as well.

I know, first world problems. Woe is me, the NAS that I hoped to have up and running is doing everything but that. To further complicate my life, I have to attempt communication with engineers.

James fucking Wang.

Update: James Wang came through, sending in a much more populated list. I’ve updated the list of qualified processors for the Supermicro X10SLL-F to reflect the new additions.

I also got a taste of Supermicro’s legendary reputed customer service when I exchanged emails in earnest with James Wang. He states that their lab will qualify CPUs on request. 🙂

RE: “Inside the 9/11 Museum”: The Challenges of Commemorating a Tragedy

For those who say rebuilding the towers would have been better, I wholeheartedly agree. I would still have put a memorial in somewhere, but I think that putting up two larger towers would have been a much more symbolic pair of middle fingers than invading two countries and turning our nation into a secret police state.

Andrew Mechem on wired.com

I was starting on the August issue of Wired Magazine when I came across Andrew Mechem’s comment, highlighted by the editors. It raises a good point about the War on Terror.

In my mind, America has always been a Christian nation. I know now that this idea actually forms a backdrop for some pseudo-legal discussion (RationalWiki on The United States as a Christian Nation), but the fact remains that the US is predominately Christian:

Protestant 51.3%, Roman Catholic 23.9%, Mormon 1.7%, other Christian 1.6%

(Source: CIA World Factbook on the United States)

I take issue with how blatantly un-Christian our nation’s violent response to the events of 9/11 has been.

Matthew 5:39, NIV: But I tell you, do not resist an evil person. If anyone slaps you on the right cheek, turn to them the other cheek also.

By exacting such terrible retribution against those responsible for the attack, we brought trouble upon ourselves. The war’s financial and human cost has been high, nearing on absurd (http://en.wikipedia.org/wiki/War_on_Terror#Costs, http://en.wikipedia.org/wiki/War_on_Terror#Casualties). Our individual liberties grow increasingly restricted in the name of preserving national security. Meanwhile, the nation’s war machine has benefitted by continued exposure to action.

There is a high price that we pay to wage war, one that is not borne by those who elected to begin it. Is it reasonable to believe that our War on Terror can eradicate all terrorist activity?

When I took AP US History in my sophomore year at the International School of Kuala Lumpur, I was fascinated by our nation’s history of armed conflict. Looking back, I feel it would be worthwhile to study the Korean War and the Vietnam War in greater depth. It seemed that we glossed over the Korean War, though I may be wrong.

Back then, I was confident that the US had lost its moral high ground in the Vietnam War. Generations of students of US History moving forward will look back on the War on Terror as the true point in which the US lost its moral high ground.

Further reading: Losing the Moral High Ground: The US and the Rule of Law

My First DIY NAS Build: Troubleshooting the Supermicro X10SLL-F + Intel Pentium G3258

OS: FreeNAS
CPU: Intel Pentium G3258
Motherboard: Supermicro X10SLL-F (~$170, Amazon.com)
Memory: 2*8GB Crucial DDR3-1600 1.35v DR x8 ECC UDIMM 240 pin, CT2CP102472BD160B (~$170, Amazon.com)
Case: Fractal Design Define R4 Blackout Edition
PSU: Corsair CS450M 450W Modular

I encountered problems soon as the build was completed. IPMI showed that the system was addressable, but I couldn’t read any values from the sensors (all sensors displayed N/A). No video output from either the onboard VGA or through an add-on video card. No beep codes from the motherboard to report.

I spoke with Supermicro technical support to learn that my board was shipped with BIOS revision 1.1. At the time that the Supermicro X10 series was released, the selection of Intel LGA1150 processors was more limited than it is now. The Intel G3258 is a Haswell Refresh (aka Devil’s Canyon) processor, so it was possible that changes (to the microcode?) broke compatibility with the Supermicro X10SLL-F. The newest BIOS revision 2.0 is X10SLL4.424, which was released on April 24, 2014.

I had two options available to flash the BIOS myself. The two options for flashing a Supermicro motherboard have been written about at length (see Bhargav’s IT Playground – update BIOS on Supermicro Servers with DOS bootable), so I’ll only share some questions and answers that I got from talking with Supermicro technical support staff.

My first conversation with Supermicro technical support led me to believe that it was possible to flash the X10SLL-F without a CPU installed. The guy on the line told me that I could purchase Supermicro SFT-OOB-LIC, a ~$20 license that enables an administrator to flash the BIOS through IPMI. I weighed my options. I could:

  1. Remove the Supermicro X10SLL-F from its case, package it, and ship it to Supermicro for them to update the BIOS
  2. This option has a turnaround time of one week, but it has a low dollar outlay (not accounting for time spent removing, packaging, shipping, and reinstalling the motherboard)

  3. Flash the BIOS through IPMI after purchasing Supermicro SFT-OOB-LIC for $20
  4. This option would get me up and running in the least amount of time, provided that the compatibility issue was only in the BIOS

  5. Take my NAS box over to Microcenter, ‘borrow’ a processor from them, flash the BIOS, and return the CPU
  6. The most unattractive option in terms of time and ethics. If I can avoid two roundtrips and forcing a retailer to sell back an item as an Open Box Item, then I would like to avoid it

I decided that flashing the BIOS through IPMI would be the best option given the circumstances. I was prepared to make the purchase of the license, but then I thought that I had better call Supermicro technical support to be sure that the IPMI BIOS update would work as I expected. A new voice greeted me on the other side of the line. When I explained what I was about to do, the man told me that this wasn’t possible. How could I flash a BIOS without a working CPU installed, he asked?

I looked at the time and decided that I had best get myself over to Microcenter. I brought my NAS with me and took it into Microcenter’s technical support department. The man working that department greeted me: I explained my situation, and he told me that I could use the room.

The Houston Microcenter has a modest selection of current-generation processors. The only one that I was confident would fit my billing was an Intel Xeon E3-1240 v3 (Intel), which I picked up for $249.99 BTAX. It was released in Q2’13, sufficiently long ago that there shouldn’t have been any compatibility issues arising from the X10SLL-F’s BIOS.

Back in the room, I opened up my MacBook Air and fired up Windows 7. I created a bootable USB flash drive using Rufus (Rufus – create bootable USB drives the easy way, free download) and copied over the files from X10SLL4_424.ZIP. I swapped out my Intel G3258 for the new Xeon E3-1240 v3, plugged in the USB drive, and booted up the NAS box.

Unfortunately, flashing the BIOS didn’t fix the issues that I had. The Intel G3258 still wouldn’t POST. I would have liked to stay and troubleshoot longer, but Microcenter was closing. I packed my things and headed home, determined to fix the problem at the start of the new day.

I swapped back to the Xeon E3-1240 v3 and went into the BIOS. It clearly showed revision 2.0, so what was going on? I ran the flash utility once more, hoping that I might gather some clues that I missed the first time.

Bingo — I saw two messages that had puzzled me the first time. AFUDOSU.EXE not found and Error when sending Enable Message to ME !!. Armed with this new information, I called Supermicro technical support once more.

I explained to the man on the line that I was encountering some issues with my X10 series motherboard. He listened for a while, and then stated that he’d try to get me in touch with the engineer that had been working on this particular board. I left a call-back number and waited.

I spoke with Milton Cai. He listened to my situation and then checked the Intel Pentium G3258 against the list of qualified processors for the X10SLL-F. It didn’t appear on the list.

Milton was kind enough to send over the list of qualified processors for the Supermicro X10SLL-F. I’m a little puzzled why Supermicro does not publish this list on their website, but it’s reproduced for you here.

Lesson learned: always consult the manufacturer of the motherboard to be sure of a processor’s compatibility before falling in love with a processor

List of Qualified Processors for Supermicro X10SLL-F

I received this (partial) list of qualified processors for the Supermicro X10SLL-F motherboard from Milton Cai, an engineer at Supermicro who has worked on that board.

Update: this list has since been updated with additional information provided by James Wang of Supermicro. I’ve expanded the information here to include processor launch dates and Pass/Fail for those processors that have it.

CPU Launch Date Pass/Fail
E3-1220 v3 Q2’13 Pass
E3-1220L v3 QF9S Q3’13
E3-1220L v3 QE9Y Q3’13 Pass
E3-1225 v3 Q2’13 Pass
E3-1226 v3 Q2’14
E3-1230 v3 Q2’13 Pass
E3-1230L v3 Q2’13 Pass
E3-1240 v3 Q2’13
E3-1240L v3 Q2’14 Pass
E3-1245 v3 Q2’13
E3-1268L v3 Q2’13
E3-1270 v3 Q2’13 Pass
E3-1271 v3 Q2’14
E3-1275 v3 Q2’13 Pass
E3-1276 v3 Q2’14 Fail
E3-1280 v3 Q2’13 Pass
E3-1285 v3 Q2’13
E3-1285L v3 Q2’13 Pass
E3-1281 v3 Q2’14 Pass
E3-1286 v3 Q2’14 Pass
i3-4130 Q3’13 Pass
i3-4330 Q3’13 Pass
i3-4340 Q2’14 Pass
i3-4360 Q2’14
G1820 Q1’14 Pass
G1820T Q1’14 Pass
G1830 Q1’14 Pass
G3220 Q3’13 Pass
G3220T Q3’13 Pass
G3420 Q3’13 Pass
G3420T Q3’13 Pass
G3430 Q3’13 Pass

The Original List from Milton Cai

G3220
G3220T
G3420
G3420T
G3430
i3-4130
i3-4130T
i3-4330
i3-4330T
i3-4340