Potential Deal: 2 x Dual 2011 nodes @$199, Quanta Openrack

tbkdan

New Member
May 2, 2016
19
2
3
32
i swapped the ram and coin battery. still the same. the good node is working, the bad node is not starting. now the good node has new problem: after about 5 minutes of post, the fans turn into high mode, very very loud, like jet taking off. no clue.
Try different RAM, did the trick for me. See below. I'm now running Kingston sticks from Natex with zero issues.

.....
  • For those who are stuck with full-speed fan issues, try different RAM. I purchased supposed "Samsung" RAM from a seller in China. This reported rock solid DIMM temps of 80C - after a cold boot. A buddy's Quanta unit reported actual temps that fluctuated with load as one might expect. Additionally, when viewing "sdr list" output via IPMI, "TSOD SMBus Sts" (TSOD = Thermal Sensor On-Die) would report a warning which is what would eventually ramp the fans up to full speed. I swapped out for a stick of desktop non-ECC DDR3 as a quick test and the system booted with zero warnings and the fans stayed silent. A workaround for silent fans is to disable the ME components in BIOS which keeps the fans from ramping up to 17kRPM, but it also kills your Serial over LAN (SOL) capabilities
.....
 

homestudio4k

Member
Apr 28, 2016
46
0
6
20
thanks, tbkdan. i actually have claimed samsung for both of my nodes each for 128gb. i will try to disable ME in bios first, then try some non ecc desktop ram later.
donjon: i did press the red button on the node and both nodes powered on.
 

grfxlab

Member
Apr 6, 2016
62
12
8
51
thanks, tbkdan. i actually have claimed samsung for both of my nodes each for 128gb. i will try to disable ME in bios first, then try some non ecc desktop ram later.
donjon: i did press the red button on the node and both nodes powered on.
Try using only two sticks of ram. Once i had a node working with two, i swapped out 1 of the working one with 1 from a group of unknown working condition. When the node booted i turned it off and tried again. I endrd up with 4 bad sticks. When 1 bad stick is installed the node powers on and the blue light stays on but it does not post or enter bios. Also disconnect your lan cable and hard drive. Just have graphics card, 2 stick mem and cpus. The fewer things connected the easier it is to test. Add things back after you get the node to post.
 

Nnyan

Active Member
Mar 5, 2012
124
32
28
sooo, anyone work on rehousing these into a standard chassis?
I'm really tempted by these, I really just need a single node though. The only thing holding me back is the power issue but I'm guessing with the converter it seems pretty simple (I do NOT solder).

I see two current offers for single nodes but they don't seem to include the power supplies.
 
Last edited:

StammesOpfer

Active Member
Mar 15, 2016
382
126
43

homestudio4k

Member
Apr 28, 2016
46
0
6
20
In my case, I tried non-ecc ram, full fan speed. Later I found out that airshroud, as previous pages mentioned, is the cause. When I swapped the ram sticks and bios batteries, I did not put on airshroud for convenience, thus the full speed fan noise. I think motherboard has some sensors for airshroud to automatically control fan speed. Anyway, in my case, airshroud is the issue. Luckily I did not buy another one from eisis, which does not have airshroud in the package.
For another node not showing video info, I tried to reconfigure jumper J46 to 2-3 and then to 1-2 to reset bios, still not working. I bought a low end video card Dell NVIDIA Quadro NVS 290 256MB PCIex16 Video Card recommended by grfxlab. Probably my GT620 is not early enough. Thanks for the forum members' help.
 

Nnyan

Active Member
Mar 5, 2012
124
32
28

LastKookie

New Member
Apr 20, 2016
3
0
1
25
In my case, I tried non-ecc ram, full fan speed. Later I found out that airshroud, as previous pages mentioned, is the cause. When I swapped the ram sticks and bios batteries, I did not put on airshroud for convenience, thus the full speed fan noise. I think motherboard has some sensors for airshroud to automatically control fan speed. Anyway, in my case, airshroud is the issue. Luckily I did not buy another one from eisis, which does not have airshroud in the package.
For another node not showing video info, I tried to reconfigure jumper J46 to 2-3 and then to 1-2 to reset bios, still not working. I bought a low end video card Dell NVIDIA Quadro NVS 290 256MB PCIex16 Video Card recommended by grfxlab. Probably my GT620 is not early enough. Thanks for the forum members' help.
there isn't any sensor for air shroud, the reason your fan is going crazy without it is because your CPU is getting too hot. the air shroud helps directing air through your CPU to keep it cool
 

homestudio4k

Member
Apr 28, 2016
46
0
6
20
the effect of the air shroud is amazing. my good node runs overnight without an issue. still quiet with centos 6.8 just installed. centos 7.2 had issue with video driver and cannot be installed.
 

voxadam

Member
Apr 21, 2016
107
14
18
Portland, Oregon
A couple quick questions:
  • What, if anything, stands in the way of using a modern GPU (e.g. GTX980) in one of these machines? Aside from the riser issue.
  • Are these machines capable of running E5-2700v2 CPUs?
  • Does anyone know if the firmware, particularly for the ME, is signed? I'd love to see coreboot running on these but if the images are signed that's never going to happen.
Thanks. I've enjoyed following this thread.
 
  • Like
Reactions: crotaphiticus

homestudio4k

Member
Apr 28, 2016
46
0
6
20
1. nobody posted results here with 980. if you experiment, you need an extra power for the card.
2. i think so. but others please confirm.
3. not sure. i don't see it in the bme. but maybe hidden somewhere else.
 
Last edited:

tbkdan

New Member
May 2, 2016
19
2
3
32
Are these machines capable of running E5-2700v2 CPUs?
The OCP server spec sheet doesn't mention anything about those v2 CPUs, nor does the C602 chipset compatibility page (only 16XXv2).
Does anyone know if the firmware, particularly for the ME, is signed? I'd love to see coreboot running on these but if the images are signed that's never going to happen.
The Quanta BIOS downloads include the full BIN (which includes the ME firmware) that can be flashed via SPI to a brand new BIOS chip (I had to do this when I got a bit too adventurous with inband flashing >.>). I don't know enough about what checks are in place, but I know you can always install a fresh BIOS with ease using the proper tools :)
 

voxadam

Member
Apr 21, 2016
107
14
18
Portland, Oregon
Thanks all for the replies.

Justin, you right of course. What I should have asked is if Quanta or Wynn configured the machines in such a way that the ME could be disabled entirely. If I remember correctly, requiring a functional (and signed) ME for boot is actually at the ODM's discretion. The ODM sets this requirement by setting fuses in the PCH. While I may have issues with the immutability of the ME it's actually not my primary concern here, I'm far more curious about the possibility of porting coreboot to this hardware and that's in part dependent on running firmware which isn't signed by a trusted Intel partner.

As far as the possibility of running v2 CPUs is concerned, I only ask because I've seen plenty of other C602 based boards on the market with such support. It wasn't until the v3 CPUs that there was a major break in backward compatibility which resulted in the "2011v3" socket.

I should have been a little less specific when I mentioned the GTX 980. While it would be ideal I'd be happy to hear any success stories of people using modern GPUs in these systems. Also, the requirement of a secondary power source for such a card shouldn't be an issue. My thought is if I use an HP Common Slot power supply with a sufficiently high rating I should be able to pull the additional 12 volts from it, if that doesn't work I can always dedicate a second PSU to the GPU.

There was one more thing that I forgot to ask in my original post. Is there any mention of "NTB" or "non-transparent bridging" anywhere in the firmware config? I'm curious because there's a very brief mention of NTB in the specification (§4.5) and while I no immediate need for it I do think it could be fun to experiment with. CPUs of this vintage all suffer from a fairly serious bug[1] but apparently a workaround is possible, but only in firmware (and now this post has gone full circle).

Again, I really appreciate all your replies, and this thread in general.

--adam

[1] Intel errata BT118, "Complex Platform Conditions during a Transition to S4 or S5 State May Result in an Internal Timeout Error"
 
Last edited:

tbkdan

New Member
May 2, 2016
19
2
3
32
Thanks all for the replies.

Justin, you right of course. What I should have asked is if Quanta or Wynn configured the machines in such a way that the ME could be disabled entirely. If I remember correctly, requiring a functional (and signed) ME for boot is actually at the ODM's discretion. The ODM sets this requirement by setting fuses in the PCH. While I may have issues with the immutability of the ME it's actually not my primary concern here, I'm far more curious about the possibility of porting coreboot to this hardware and that's in part dependent on running firmware which isn't signed by a trusted Intel partner.

As far as the possibility of running v2 CPUs is concerned, I only ask because I've seen plenty of other C602 based boards on the market with such support. It wasn't until the v3 CPUs that there was a major break in backward compatibility which resulted in the "2011v3" socket.

I should have been a little less specific when I mentioned the GTX 980. While it would be ideal I'd be happy to hear any success stories of people using modern GPUs in these systems. Also, the requirement of a secondary power source for such a card shouldn't be an issue. My thought is if I use an HP Common Slot power supply with a sufficiently high rating I should be able to pull the additional 12 volts from it, if that doesn't work I can always dedicate a second PSU to the GPU.

There was one more thing that I forgot to ask in my original post. Is there any mention of "NTB" or "non-transparent bridging" anywhere in the firmware config? I'm curious because there's a very brief mention of NTB in the specification (§4.5) and while I no immediate need for it I do think it could be fun to experiment with. CPUs of this vintage all suffer from a fairly serious bug[1] but apparently a workaround is possible, but only in firmware (and now this post has gone full circle).

Again, I really appreciate all your replies, and this thread in general.

--adam

[1] Intel errata BT118, "Complex Platform Conditions during a Transition to S4 or S5 State May Result in an Internal Timeout Error"
Again, I'm not sure about signature verifications, but you can disable ME via the BIOS (check page 95 of the tech manual) and via a jumper (for forced BIOS upgrades).

My GTX950 would start booting enough to show the splash screen but then just hang there. Switching to an older AMD worked fine. That said, I was using the included risers which split the x16 into two x8 so that may have caused issues too.
 
  • Like
Reactions: voxadam