|
2016-05-21, 14:12
|
#34 | ||||
|
Dec 2014
111111112 Posts |
I am a software guy and not a hardware guy. But after the IO port conflict I
started wondering about getting to the 8 th GPU also. I switched the disk from AHCI to IDE but it seems to be using more IO ports. Quote:
If I disable USB will the keyboard still work for the BIOS? Some C6100 use PS/2 keyboard and mouse, but mine uses USB. Quote:
Quote:
Quote:
|
||||
|
|
|
2016-05-21, 14:14
|
#35 | ||
|
Dec 2014
3·5·17 Posts |
Quote:
Quote:
|
||
|
|
|
|
2016-05-21, 14:34
|
#36 | |
"/X\(‘-‘)/X\"
Jan 2013
https://pedan.tech/
24×199 Posts |
Quote:
echo 1 | sudo tee /sys/bus/pci/devices/0000:10:00.0/remove Last fiddled with by Mark Rose on 2016-05-21 at 14:34 |
|
|
|
|
|
2016-05-21, 15:05
|
#37 |
|
Dec 2014
3×5×17 Posts |
The ipmitool returned some errors.
The sticker on the top front of the case does say version A00 like yours. Here is a screen shot of the web interface showing "8 to 1" configured. Notice the picture says "Control by O Jumper X BMC" which leads me to think there might be jumpers inside the case to configure "8 to 1" mode. |
|
|
|
|
2016-05-21, 15:48
|
#38 | |
|
Dec 2014
3·5·17 Posts |
Quote:
I tried a couple of the rescan and one remove. nivida-smi gave no joy. Removed and restarted the nvidia driver, and still no joy. Since I do not know what I am doing here, I only tried a couple. |
|
|
|
|
|
2016-05-21, 21:23
|
#39 |
|
2·5·19·43 Posts |
As for USB, I think the USB keyboard will stop working if you disable the ports. Do you actually need to use the server in person with a keyboard routinely and are you just concerned about getting back into the BIOS again later? If the only concern is the BIOS, perhaps a better option is to try disabling USB just in linux - disabling it in the BIOS will improve the chances of it working, as then the BIOS has an improved chance of assigning the memory correctly, but it is easier to try with Linux alone. To do this, press e to edit when at the grub menu and then add the following to the list of kernel parameters:
nousb This takes effect for the current boot only, and will go away when you boot again. There is a way to make it permanent, if it helps. If you were to disable USB in the BIOS, the answer to recovering keyboard input later would be to see if you can edit the BIOS settings within Linux (say via ssh connection) using Dell's setupbios tool, available here: http://poweredgec.com/files/ And described here: http://www.dell.com/support/article/us/en/04/SLN266215 I would suggest testing this first to make sure it works, before disabling USB and then needing to rely on it. Another and better alternative (in case of any Linux problems occurring) is to set up serial console redirection via IPMI so that you can control the keyboard input remotely, but you would need to be very careful with setting this up, as there are apparently bugs with some firmware versions on the C410X that could cause problems and even brick your board, as described here: https://forums.servethehome.com/inde...oard-yet.1448/ In the worst case, you could unplug the power, pull out the node, remove the CMOS battery, and use the reset jumper on the motherboard, and that should reset the BIOS to get USB working again. Thinking of firmware, have you checked your BIOS version on the C6100? An update might fix the problem. However, there is of course always some risk, as the board can be bricked if a problem occurs during the update. You apparently also have to be careful to ensure that you don't have a custom C6100 that doesn't work with the regular BIOS, as described here: http://www.copyerror.com/2013/08/10/...ll-c6100-bios/ |
|
|
|
2016-05-21, 21:25
|
#40 |
|
2×17×89 Posts |
Sorry - I should have explained in more detail what I was talking about with the PCI stuff. A BAR is a base address register, i.e. a region of memory on the PCI card that needs to be mapped into the computer's memory or I/O ports. The dmesg output from earlier showed that one of the BARs was not being mapped correctly.
With the bus remove and subsequent rescan options I gave, I didn't explain clearly enough that they must be carried out in order: i.e. run the remove command, then the rescan that I placed on the next line - i.e. do not run the rescan on the device that you removed, but the bus device. Ideally, also run this rescan: echo 1 > /sys/bus/pci/devices/0000:0e:04.0/pci_bus/<whatever>/rescan where you will need to look to see what whatever is - there should only be a single entry there. Essentially, the idea is to remove the GPU from the bus, then rescan the bus to try to pick it up again, where the bus in this case means a bridge chip in the C410X far down the tree, so that you don't end up rescanning every other GPU too. Unfortunately the tree output isn't completely clear because the extra spacing was removed and so it is difficult to work out where the | lines should go to show the connections between bridges. Please could you attach the output from lspci -tv as a text file? Please can you tell me what the error was with the ipmitool command and can you try again with the following - it might help: ipmitool -I lanplus -H x.x.x.x -U root -P password fru print Again, please replace password with your actual password and replace root with your actual username (if you changed it), and of course x.x.x.x with the IP address. Just to confirm, was the IP definitely set correctly to the c410x's IP and can you reach the web browser interface on the C410X from the computer on which you can ran the command? After looking at various boards I have seen for sale, they all only seem to have jumpers for switching between 2:1 and 4:1, so I think 8:1 is a software setting, making it difficult to determine which boards support it form visual inspection. I think that the chassis and board revisions need not be the same, so it would be really helpful for me to know if you have a Rev A00 board, A01, or even higher, along with the manufacturing date. I think you might have a Rev A01, even though the chassis is Rev A00. Finally, just in case it is useful to you, did you know that straightforward control of the c410x is available from the command line, without needing to use the web browser? You can download C410x_Scripts.tar.gz and find inside a file called port_map.sh, which can be used to change the mapping from the command line. You can get this file from: http://poweredgec.com/latest_powered...stem Utilities (don't get a Scripts file version starting with an R - they don't support 8:1 mode). Then there are also various power commands, including individual slot power commands, that you can see here: http://www.dell.com/support/article/us/en/04/SLN244176 |
|
|
|
2016-05-22, 02:01
|
#41 | |
|
Dec 2014
111111112 Posts |
Here are the IPMI errors
Quote:
|
|
|
|
|
|
2016-05-22, 02:06
|
#42 |
|
Dec 2014
3·5·17 Posts |
lspci -tv output
|
|
|
|
|
2016-05-22, 02:43
|
#43 |
|
Dec 2014
3·5·17 Posts |
I opened up the box and took a picture of the PCB.
It has a sticker that says A00 but the PCB says "rev 1.1". I will leave it open in case you want more pics. |
|
|
|
|
2016-05-22, 02:54
|
#44 |
|
5×997 Posts |
Thanks! That's very useful and kind of you to do that. Please can you also tell me the revision from the iPass connector cards, both on the sticker and on the PCB? In particular, the upper card is likely to be relevant.
|
|
|
 |
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Running multiple ecm's | johnadam74 | GMP-ECM | 21 | 2019-10-27 18:04 |
| Multiple GPU's Windows 10 | cardogab7341 | GPU Computing | 5 | 2015-08-09 13:57 |
| Using multiple PCs | numbercruncher | Information & Answers | 18 | 2014-04-17 00:17 |
| Running multiple copies of mprime on Linux | hc_grove | Software | 3 | 2004-10-10 15:34 |
| Multiple systems/multiple CPUs. Best configuration? | BillW | Software | 1 | 2003-01-21 20:11 |
