Supermicro mbd-x9scm-f-o

[nyquist]

Hey guys,
After reading some reviews and forums for a couple of hours about the SUPERMICRO MBD-X9SCM-F-O motherboard, I've founds some people who had bad experiences with this mobo, in the past..

They were kind of beta testing it when it was released and I understand some enhancements were made by Supermicro in order to fix all the problems (e.g.: different intel network chipsets were added to fix some compatibility issues).

Beyond that, I've found some threads at forums.freenas.org from people who was taking in mind this mobo for their builds, so I was looking for some thumbs up or down over this choice.

The reason why I'm asking for this specific mobo is because it's the only "easily" available Supermicro motherboard in my country.

Thanks in advance,

 

[9C1 Newbee]

I have had this board running for about a month. I have nothing bad to say about it. I give it a thumbs up.

 

[jgreco]

The compatibility issue is most likely that the X9SCM (C204) and X9SCL (C202) use an 82579LM for the primary ethernet controller. While an awesome ethernet controller, it is too new to be supported out of the box by ESXi4.* and ESXi5.0, and by a lot of other operating systems. So it winds up very frustrating because after you add the necessary driver, if your OS has already wired its ethernet configuration in, you end up with the network ports being backward when the driver for the 82579 gets installed and shows up as the second ethernet.

Supermicro ended up releasing the X9SCL+ as an appeasement to all their exasperated virtualization customers, the only meaningful difference being a pair of 82574's instead of the 82574/82579 combo.

This isn't a problem for FreeNAS, and my estimation is that the 82579LM is a slightly better choice for ethernet, but really even the 82574L (technically more of a desktop ethernet controller, used in the Gigabit CT for example) is pretty awesome under FreeBSD. However, the X9SCM has more PCIe capacity and the option to do SATA 6Gbps on two ports so it may well be the best general FreeNAS option out of all the Supermicro X9S* boards. That's why I specifically call it out in the sticky.

If you plan to do ESXi on your board at a future date, picking the X9SCL+ instead will save you some minor frustration and inconvenience, but that's really the only reason to do that that I can think of. We run X9SCL+'s and X9SCi's here with great success, but we're an ESXi shop.

 

[9C1 Newbee]

I guess I should have added a little more information. I am running FreeNAS-8.3.1-RELEASE-p2-x64 on the board.

I will also add, I had a bit of trouble getting the ethernet controllers to work. I migrated my FreeNAS configuration file from the old machine to the new one. Doing that is known to cause all kinds of issues. Once I started FreeNAS fresh the issue went away. Again, that was no fault of the board.

 

[nyquist]

Guys, thanks for sharing your experiences. I'll be posting my final build choice for a proper revision, soon.

Regards,

 

[Hugo Ochoa]

I have this same motherboard (see my sig). After building the server using a Corsair TX750w power supply, I see the fan's try to start spining and then nothing happens. The system does not start at all. If I push the power button again, the fans try to start but not even the power suply fan starts. The standby and PMI LEDs turn on (PMI blinks slowly as it should) and the onboard battery is doing 3V. I've disconnected all but CPU, one stick of RAM on DIMM 2A, and power switch and still nothing. I've tried switching the RAM sticks arround and even re-seated the CPU. The power supply is the only part that is not brand new as it was removed from a system that was working fine and it is 80+ certified with plenty of power for this build. I'm now thinking that the motherboard is DOA and will need replacement.

Does anyone have a clue what could be wrong here? This is not my first buid and as always I followed strict anti-static/clean-room procedures while putting it together. This is however, my first SuperMicro build and I have to say I'm not very impressed with the looks of the board. It just doesn't look as well built as other boards I've used by diferent manufactures such as Asus or Asrock.

 

[ser_rhaegar]

You should probably start a new thread for help on this issue Hugo.

 

[jgreco]

It just doesn't look as well built as other boards I've used by diferent manufactures such as Asus or Asrock.

If by "well built" you mean that it lacks enthusiast bling, yes, you're correct. The difference is that the SuperMicro board is designed for servers, and they're not show-offy, just extremely well built (for values of "well built" that include "very likely to be running five years later") and very functional. You won't find another quality manufacturer with such a breadth of server offerings.

 

[Hugo Ochoa]

If by "well built" you mean that it lacks enthusiast bling, yes, you're correct. The difference is that the SuperMicro board is designed for servers, and they're not show-offy, just extremely well built (for values of "well built" that include "very likely to be running five years later") and very functional. You won't find another quality manufacturer with such a breadth of server offerings.

I appreciate your response jgreco. I'm not referring to the color or shine of the motherboard. I've never bought a motherboard based on its looks. I'm referring specifically at the look of the solder on components of the board. Also, I noticed a certain amount of give when I plugged in the VGA connector. Looking at other motherboards made by different manufacturers, I see that the solder is not as neat and well done. I can see that the looks of it may not have anything to do with its quality and durability. But it makes me suspicious nonetheless. It may be an unfair assessment on my part since I'm comparing this server motherboard to PC motherboards that are meant for overclocking and can work at very high temperatures. I understand building a PC is not the same as building a server, but it’s pretty close to the same process. I’ve never had any other motherboard before with this problem. Do you have any suggestions on the problem I’m having?

 

[jgreco]

I'm not a fan of the new super-thin VGA connectors but I've seen similar "small format" connectors appearing elsewhere. They do have an uncomfortable amount of play. I was rather annoyed to discover that starting sometime around X9, the hex nuts on those were no longer removable, which meant fastening the connector to the backplate was no longer possible, and it meant that some I/O shields needed to be modified accordingly.

I haven't noticed any issues with the quality of solder on the boards, but I will note that across the industry RoHS was a substantial "hit" to the strength of solder joints and has transformed what was once not a particularly difficult area of manufacturing and rocketed it into "wow that sucks."

http://www.ecnmag.com/articles/2011/12/was-lead-free-solder-worth-effort

It is possible that you are just noticing a different strategy being used.

FWIW we still use lead solder in our shop

 

[cyberjock]

I'll second the RoHS. That has ruined long term reliability for a lot of electronics. Luckily, for fields where electronic reliability is couple to life/death situations(like nuclear power plants) we still can buy non-RoHS stuff from some companies(but now it's super expensive because of the "feature"). My understanding is that many military and government contracts(think NASA) pay lots more for non-RoHS stuff since you have a dire need for it to work reliably for years to come.

Not that I disagree with you Hugo Ochoa, but I do find it somewhat entertaining(in a non-offensive way) that you are complaining about soldering jobs. Anyone who worked in electronics fields saw a sudden rapid decline in quality the second RoHS was pushed on the world.

 

[9C1 Newbee]

I never knew what a RoHS was. I see the symbol everywhere. Now I understand. This is more along the lines of the mandatory low flow toilets you have to flush 4 times. I think this is just another excuse to move away from quality. Less to do with the environment and more to do with $$$.

 

[cyberjock]

I never knew what a RoHS was. I see the symbol everywhere. Now I understand. This is more along the lines of the mandatory low flow toilets you have to flush 4 times. I think this is just another excuse to move away from quality. Less to do with the environment and more to do with $$$.

No. It has to do with people that fail to see the trees in the forest. They focus on one single thing and try to solve that problem to the exclusion of everything else. Example:

So pretend my child ends up mentally handicapped because of lead poisoning. Somehow I've decided it's because I work on PCB boards all day. So I start this initiative to get lead out of solder because lead is very unhealthy. Well, let's say I manage to convinced the whole damn planet that lead-free is better for the environment. After all, if lead is bad then no lead must be better, right?

But, I totally failed to factor in the fact that far far more electronics will end up in the landfill sooner than expected because a certain someone didn't want lead in their solder.

Another example:

So pretend that I hate oil companies and think they are evil and that we all must do our part for the environment.. And since I own a farm and sell massive quantities of corn, why not put both together? So I'll convince the world that ethanol should be forced into the gas pumps to help the environment since it burns so much cleaner than standard gasoline. After all, cleaner emissions must be better, right?

But, I totally failed to factor in the simple fact that your average gasoline engine will actually have to burn a larger quantity of fuel to get the same amount of power out. I also didn't factor in the actual energy costs to manufactuer said corn-based ethanol.

------------

So yeah, all a big scam. But a few morons are convinced they did the world a favor.

 

[Hugo Ochoa]

I got the replacement motherboard from Newegg and installed it with a minimal amount of components just for testing. CPU, 1 stick of RAM in DIMM 2A (followed the motherboard manual as close as I could). I'm getting the system to post!!! However, the CPU fan doesn't start and I'm very worried about that. Is this normal? Does it start when the temperature of the CPU reaches a certain value?

I haven't let the system run more than 30 seconds because I'm worried the CPU fan (stock Intel) is bad and it could fry my CPU.

 

[jgreco]

Yeah, that's kind of bad.

 

[Yatti420]

Did you forget to plug the fan in? Try another fan header and see what happens.. I'm happy with the board.. Only thing I'm having issues with is FreeNAS - IPMI integration..

 

[Hugo Ochoa]

oh man I feel like such a dumb ass. the CPU fan was connected but when I did that, I pulled the wire too hard and it was stopping the fan from turning. it's all good now (I hope)! running the system with all the 32GB of ram and starting on the hardware testing. Can anyone recommend a CPU stress test in FreeBSD that I could run to make sure the CPU didn't get damaged when the fan was stuck?

 

[jgreco]

cpuburn? burnBX and friends?

 

[jgreco]

Oh and as long as the heatsink was attached correctly and it was only 30 seconds, you ought to be fine.

 

[9C1 Newbee]

Although not in free BSD, I recommend Hirens Boot CD. It is a self booting CD that can be put onto a thumb drive. It has a boat load of utilities including an array of mem test and cpu burn in utilities. Also has WDIDLE3. Might be worth a look.

http://www.hiren.info/pages/bootcd