C2SEE woes - what am I doing wrong? Give me other alternatives

[directo]

Is anyone else using this board?  I was the one that wrote a few weeks back about my problems with the C2SEE ( http://lime-technology.com/forum/index.php?topic=3520.0 ).

 

I got another RMA board back and am having real problems with this guy.

Here's some things I tested:

 

1) I decided to start with the bare minimum in the chassis - nothing hooked up!  Seems to boot and POST fine, can get into the bios.

 

2) When I add a SATA CDROM drive - seems to work fine (can boot from my windows XP disk that's in there.)

 

3) Once I add a HD (tried a WD10EADS, and a Samsung drive (the model escapes me right now)) it just crashes and powers down.

 

4) When I try to boot from JUST the unRAID USB stick, it will decompress the image, and then shutdown (bzimage / bzroot).  I've also tried Ubuntu - same thing.

 

5) I've managed to do the first part of the XP install (formatting and copying files) to the WD10EADS but when it restarted, it got to the Windows XP screen for (probably a millisecond) then shut down.  I'm trying this again with another HDD to see if it's the drive (I don't think it is, but I want to make sure).  [AGGHH - just did the exact same thing using the laptop HDD - it won't even boot into an OS!!!]

 

I've tried calling Supermicro for tech support, but they weren't too helpful, and told me to reinstall Fedora 7 or Windows on the SATA HDD (which is why I'm doing the above).

 

Is anyone else having problems with this board?  I would find it really hard to believe 2 boards came to me bad.  (I've also tested the memory using the unRAID stick and let it run over night).

 

All my components (save the memory) have been used on another motherboard without issues.  I'm really at a loss for what could be wrong.

I'm also using a Corsair TX750W powersupply, and an Intel Celeron 1.6 GHz 420.

 

How can this be the reference board?

[SSD]

Sorry you are having problems, and certainly appreciate the frustation you are having.

 

Your problems may be the fault of the C2SEE, but may not.  This IS the board that LimeTech uses in their servers, and a number of users here have used it with success.  You can be confident that if the board is not defective, that it is compatible.

 

1.  As I read your post, the first thing that occurred to me is that it may be your memory.  When you boot the USB stick, one option should be to run a memory test.  I'd suggest trying that.  If memory is bad, it can cause all sorts of symptoms and problems.  "Fancy" memory often requires a voltage boost, and some motherboards don't provide the BIOS feature to provide that boost.  You might want to post your memory brand and model to double check.

 

2.  Try double checking the PSU connections to the motherboard.  Unplug and replug the large connector to the motherboard, as well as the smaller (4-pin or 8-pin) connector.  Make sure it "clicks" securely into place.

 

3.  If the CPU gets too hot it will shutdown the system.  CPU overheating is caused by an improperly installed HSF.  If can also be caused if if the fan part of the HSF is not hooked up / spinning.  When you turn on the computer double check that the fan is spinning.  You should also be able to go into the BIOS and see the CPU temperature (normally their is a "Health monitoring" or similar area that will display this info).  Make sure that the CPU is staying at a safe operating temperature.

 

4. PSUs have protective circuitry to power themselves down if too much power is drawn or if something else is going wrong inside the PSU (e.g., it gets too hot).  It may be that the PSU is defective and only able to delivery a small percentage of its rated power.  Or that something is shorted out and causing the PSU to become overloaded.  For example, if the motherboard is shorting to the chasis.

 

5. It sounds like the problems start when you start to hook up drives.  I'd try hooking up just one drive at a time and see if you are able to avoid the hard power down.  If they each work by themselves, but together they cause problems it is looking more and more like a PSU issue.  If one drive works and another doesn't, that one drive may be broken.  (Although possible, I have never seen a hard drive cause this type of problem).

 

Hope this helps.  Please try to stay is "problem solving mode" rather than "angry blaming mode" while you try and figure this out.  If it turns out that you have a 2nd defective C2SEE, you certainly can be critical of the Supermicro QA procedures.  But until you know, keep plugging.  FWIW, your symptoms don't "feel" like a MB problem to me.

[WeeboTech]

 

Since you've already done a memory test

 

Have you tried a different (temporary) power supply?

 

Double check all connectors, the 20+4 pin connector plus the EPS12V (8PIN) connector.

 

Check for shorts on the mounting (although if there were shorts, it would drop right away).

 

An over heating CPU would shutdown the machine, however this would happen during memory tests too.

 

 

 

 

[SSD]

Oops - sorry missed that parenthetical about having already run the memory test.

 

I really think the PSU (or a bad connection from the PSU) is the most likely suspect at this time.

 

[directo]

Thanks for the replies...

I'm running (another) memory test overnight, just to double check.

 

I don't have another PSU with an 8 pin connector, BUT - I have been using this PSU with my former motherboard, and it's performed without any issues.  I'm reluctant to go out and buy another PSU to test this out -> what are the chances that it is the PSU manifesting this problem? (my previous board only used a 4pin EPS connector though).

 

I've also tried using another CPU on the board, and it didn't change any of the symptoms.

 

Even the tech support guy from Supermicro thought this was weird.

 

Some answers:

1) I'm using OCZ DDR3, 1066 memory.  Will rerun the memory test, but when I did it last week, didn't have any problems with my memory.

2) Reseated the connectors multiple times.  Everything was connected well.

3) Fan was working fine.  Reseated, and reapplied thermal grease, and tried 2 different CPU's.  Same results.

4) Haven't been able to try another PSU - but the PSU has worked with my other MOBO.

5) I can't even boot the Unraid, or Fedora live usb images  It gets to the point where it says decompressing kernel image, and the whole thing shuts down (no error beeps, or anything - the screen goes magenta, and then dies.

 

What do you guys think about:

1) the chances of getting 2 defective boards in a row...

2) my PSU being the problem (even though the one I'm using has been working without a problem with my previous microATX MOBO)

[SSD]

Are the symptoms 100% reproducable?  Does the crash always seem to happen at the exact same time in the boot?  Or does it get further sometimes versus other times?  Be as specific as possible about what you think it happening at the moment of the hard power down.

 

The PSU is still my first most likely culprit.  But if you have run the exact same PSU on another motherboard without any problems it reduces that chance.  Hate to have you buy a new one only to realize that isn't it.

 

The MB could be bad, but 2 bad MBs in a row seems very unlikely.  Are the symptoms of failure with this MB the same as the symptoms with the prior one?

 

I once had some weird results hooking the front panel LEDs to the wrong pins on the MB.  May be worth a try to confirm that everything is hooked up perfect.

 

Verify that the 8 pin power connector matches the pinout in the MB manual.  I have never used one and if the PSU and MB are not following the same standard, could cause problems.  (This is a long shot, but the 8 pin connector is not being used with the ITX motherboard).

 

I'll sleep on this and may have more ideas in the morning.

[directo]

- symptoms are 100% reproducible -> when trying to boot linux, it dies at the exact same time, when trying to boot from the HDD (WinXP) it dies at the exact same time.  when booting unraid, it gets through the bzimage and bzroot steps (with the dots), and then shortly after that, it shuts down.

 

- my worries exactly, I'd like to rule everything out before I try and get my hands on another PSU.

 

- as of now, yes - the symptoms are the same; however, when I tried the first C2SEE a few weeks ago, I could actually boot into unRAID but the system would crash whenever i spun a disk down, or did a file transfer.  Now I can't even boot into unRAID.

 

- have double checked all my connections, and have tried only connecting the power switch up and no comination has worked

 

- the 8 pin power connector can only go in a certain way (some pins are square, others are tapered at one end), so it's snapping in correctly

 

let me know if a video would help - i can try whipping something up and posting it on youtube or something.

 

 

[WeeboTech]

This is a tough one.  Strange.. Memory allocation, speed ?

 

In this case I would disconnect everything except the bare minimum of what is needed.

I.E. just try booting unraid from the USB stick.

 

I would erase the CMOS and start from scratch.

Load fail safe defaults.

I would disable everything in the BIOS possible. I.E. Serial, parallel, Firewire. Audio.

 

I would relax any memory timings to the slowest possible.

Turn off quiet boot, turn off quick boot.

Turn off the high performance event timer.

 

Turn off ACPI or try different versions.

I've had issues with prior motherboards with ACPI, and also when having Intel Speed Step and C state enabled simultaneously

 

Fro graphics mode I would use the least memory that can be allocated (gives memory back to the system).

I also might try a PCI video adapter and disable internal video just to see what happens.

 

USB Functions, use the least amount possible.

 

Disable the floppy drive or controller

DIsable the ide controller if not being used.

 

Hardware health I would leave on.

However I would go into the bios and watch the bios hardware health screen to see what the temp and fans show.

 

you could try slowing down the USB. I.E turning off USB 2.0.

 

If none of these work it's the memory or the power supply.

The motherboard should be able to boot into unRAID without any hardware.

 

 

Motherboard Manual here for anyone else who might have ideas

http://www.supermicro.com/manuals/motherboard/G45/MNL-1043.pdf

 

 

 

 

[SSD]

It seems that these hard shutdowns are occuring in conjunction with the first significant disk activity. Try disconnecting ALL drives (disconnect power and signal cables) and see if unRaid will boot.

[RobJ]

One last thing to try, AFTER all of the very good suggestions above, is to verify your components (PSU, cabling, RAM, etc) still work in the original working system.  There are risks moving from one system to another, pins get bent, wires come loose in connectors, other issues possible.  I know it is a hassle, but you may need to verify everything still works, after having moved them from a working system through 2 non-working systems.  That's a fair amount of pulling and tugging on wires and connectors and cards.

 

Also take a hard look at the temps and voltages on the BIOS PC Health screen, both with no drives connected and with drives connected.

[directo]

I've been trying what bjp suggests.  Just trying to boot the unRAID O/S (I've also tried Ubuntu from USB and CD).  Same problem.  (I didn't have any drives hooked in.

 

I've gone back and forth between both MOBO's a few times to make sure i wasn't going crazy.  I've also tried both C2SEE boards and same results.  I'm going to try and fiddle around with the BIOS settings (as suggested by Weebo), but may go out and get another PSU.

Any suggestions (hopefully they have a good return policy )?

 

BTW - temperatures and voltages all look ok in the BIOS.

CPU Temp - Low

System Temp  - 30-40 C

VCORE - 1.288

AVCC - 3.376

3.3Vcc - 3.376

12V - 12.249

V_DIMM - 1.584

5V - 5.152/5.184

-12V - -12.472

3.3Vsb - 3.360

Vbat - 3.280

[directo]

is there a way to boot in an attended fashion? ie. hit a key for every step?  I can see some output after the bzimage and bzroot steps...

[dandirk]

wow very puzzling...  I just setup a C2SEE and corsair 750 unraid server.  I was getting stalls in unraid boot, it wouldn't shutdown just stall at the same place each time.  For me the CPU needed to be remounted the board told me temps were "high".  This is not the case for your board though.

 

I would go out an get a PSU, you can always return it if its too expensive or doesn't work.  I would also try and isolate each part, 1 stick of ram (check each slot and each stick) etc.  Change your sata cables (though I would just get unraid to boot first, no HDD).

 

If its the PSU I almost bet your first board will work.  You may also want to look at the capacitors on the motherboards and make sure none are budging or black.  I have seen cases where a PSU has ruined a motherboard too which for me at least was a very similar experience, nothing made sense in troubleshooting until I replaced both.

 

Just so you see a little light ahead, besides my CPU seating issue the C2SEE setup was a breeze.  I didn't have to adjust much in the bios to get it up and running.  You are not crazy, it works practically right out of the box.

[WeeboTech]

is there a way to boot in an attended fashion? ie. hit a key for every step?  I can see some output after the bzimage and bzroot steps...

 

Please adjust the bios. If it's hanging after the dots from bzimage/bzroot it's not initializing hardware correctly.

 

Lower the memory timings may reveal something here. Perhaps even taking one out so it is not running in dual channel mode.

 

I've seen this before with bios and relocation issues on an old motherboard I was trying to boot from usb on a floppy.

[directo]

I guess this is how crow tastes?

 

I just went out and bought a PSU (from a big box store) to see if it works, and for now - it seems to be working (able to boot, i'm doing a transfer from 1 drive hooked up).  Previously, it crapped out by now, but I'm not holding my breath.  It's a BFGtech 550W, so I think I'll return this and get another PSU...any suggestions?

 

What's very weird is that my old PSU worked on my old MOBO (albeit, it was a uATX board, so only needed the 4pin 12V connection).  Very strange indeed.

 

Many thanks for all the suggestions - and a lesson to all, NEVER ASSUME something works just cuz it did on your old setup!

 

I'll post later on with what PSU I get, and the stability of the system after the fact....

Send any PSU suggestions along

(I'm in Canada, and will likely go to www.canadacomputers.com or tigerdirect.ca)....

[prostuff1]

Take a look at http://www.silentpcreview.com/article699-page1.html

 

They have some very good suggestions on quiet and efficient PSU's

 

 

[SSD]

Glad the mystery is solved.  PSU always seemed pretty likely to me, but without a parts supply store to help in your basement it can be hard to swap out components you don't have.  Directo, you were smart to try and rule out everyting you could without making the purchase, and also wise to realize that it was time to try a new PSU.

 

Now get yourself a good PSU and get that long anticipated array up and running.  ;D

[WeeboTech]

I guess this is how crow tastes?

 

I'm also using a Corsair TX750W powersupply, and an Intel Celeron 1.6 GHz 420.

 

I'm really surprised you had this difficulty. Could be parts failing in the power supply and the draw goes greater then it should.

Have you checked if your power supply is under warranty?

[prostuff1]

I guess this is how crow tastes?

 

I'm also using a Corsair TX750W powersupply, and an Intel Celeron 1.6 GHz 420.

 

I'm really surprised you had this difficulty. Could be parts failing in the power supply and the draw goes greater then it should.

Have you checked if your power supply is under warranty?

 

I was surprised to see that this was a Corsair PSU also.  In general they make very good stuff and I run the 550W version in my unraid server.  Check to see if it is under warranty and if it is i say exchange it.

[dandirk]

I agree with the rest... corsair is a very good brand.  They get good marks (well better then most) from HardOCP which in their PSU reviews are so tough that most PSUs fail in one thing or another.

 

I also would say check your corsair warranty, you probably can get another one from shipping costs...

[heffe2001]

I'm using a Corsair 750 single-rail in a jukebox project (needed the high 12v amperage), and it will absolutely NOT power my Intel Atom330 motherboard.  Does the exact same thing as yours was doing.  I put it in another system and it worked great there.  Had to end up getting a seperate supply for the 330, and I'm just powering the amplifier and card flip-box with the 750 (and it works great in that capacity as well).  I'm not 100% sure why, but for whatever reason that supply would cause the system to crash as soon as you tried to boot anything (either a CD/DVD, or a HD).  I was going to post this as a possible solution to your problems, but as I read down the thread saw that it was indeed the problem..

 

For the record, I'm running a Corsair vx400 in my 11-drive Unraid system, and thus-far no power issues (stacker issues, HD issues and others I've had, but nothing power related).  I'll probably upgrade the PS at some point, but I'm a firm believer in the "If it ain't broke, don't fix it" motto...

 

 

[directo]

Thanks for the update heffe - cautionary tale for those using the TX750 - it may or may not work for you. 

 

I've since upgraded to the HX620 (currently running with 13 drives).  So far so good.

 

Wanted to thank everyone who helped out!

[dandirk]

curious...  I wonder if the TX750 has issues with specific motherboards or CPUs?  Maybe the CPU is shutting things down due to voltage/current? 

 

I am going to assume maybe the latter since I have a TX750 running a Core2 on the C2SEE just fine.

 

Yet again Lime's official setup uses a celeron as well...  hmmm maybe it just non-consistent manufacture which is disappointing.