Re: [VirtualGL-Users] Dual GTX 1080i headless server issues

[<sta...@ms...>]

Hi All,
I have changed my display manager from MDM to lightDM. It was a little  
buggy to get it going but it is now finally operational and seems to  
be working appropriately.
Doing the "sanity check" tests results in this:
marshall@computer ~ $ xauth merge /etc/opt/VirtualGL/vgl_xauth_key
marshall@computer ~ $ xdpyinfo -display :0
name of display:    :0
version number:    11.0
vendor string:    The X.Org Foundation
vendor release number:    11804000
X.Org version: 1.18.4
maximum request size:  16777212 bytes
motion buffer size:  256
bitmap unit, bit order, padding:    32, LSBFirst, 32
image byte order:    LSBFirst
number of supported pixmap formats:    7
supported pixmap formats:
     depth 1, bits_per_pixel 1, scanline_pad 32
     depth 4, bits_per_pixel 8, scanline_pad 32
     depth 8, bits_per_pixel 8, scanline_pad 32
     depth 15, bits_per_pixel 16, scanline_pad 32
     depth 16, bits_per_pixel 16, scanline_pad 32
     depth 24, bits_per_pixel 32, scanline_pad 32
     depth 32, bits_per_pixel 32, scanline_pad 32
keycode range:    minimum 8, maximum 255
focus:  PointerRoot
number of extensions:    26
     BIG-REQUESTS
     Composite
     DAMAGE
     DOUBLE-BUFFER
     DPMS
     DRI2
     Generic Event Extension
     MIT-SCREEN-SAVER
     MIT-SHM
     Present
     RANDR
     RECORD
     RENDER
     SECURITY
     SHAPE
     SYNC
     X-Resource
     XC-MISC
     XFIXES
     XFree86-DGA
     XFree86-VidModeExtension
     XINERAMA
     XInputExtension
     XKEYBOARD
     XTEST
     XVideo
default screen number:    0
number of screens:    1

screen #0:
   dimensions:    640x480 pixels (169x127 millimeters)
   resolution:    96x96 dots per inch
   depths (7):    24, 1, 4, 8, 15, 16, 32
   root window id:    0x43
   depth of root window:    24 planes
   number of colormaps:    minimum 1, maximum 1
   default colormap:    0x20
   default number of colormap cells:    256
   preallocated pixels:    black 0, white 16777215
   options:    backing-store WHEN MAPPED, save-unders NO
   largest cursor:    640x480
   current input event mask:    0x420003
     KeyPressMask             KeyReleaseMask           StructureNotifyMask
     PropertyChangeMask
   number of visuals:    2
   default visual id:  0x21
   visual:
     visual id:    0x21
     class:    TrueColor
     depth:    24 planes
     available colormap entries:    256 per subfield
     red, green, blue masks:    0xff0000, 0xff00, 0xff
     significant bits in color specification:    8 bits
   visual:
     visual id:    0x41
     class:    TrueColor
     depth:    32 planes
     available colormap entries:    256 per subfield
     red, green, blue masks:    0xff0000, 0xff00, 0xff
     significant bits in color specification:    8 bits
marshall@computer ~ $ /opt/VirtualGL/bin/glxinfo -display :0 -c
name of display: :0
Error: couldn't find RGB GLX visual or fbconfig
Error: couldn't find RGB GLX visual or fbconfig

Thank you again for all the advice.

Marshall Stageberg
Michigan State University

Quoting DRC via VirtualGL-Users <vir...@li...>:

> If LightDM is available under Linux Mint (I can't recall whether it is
> or not), that would be the next thing to try.
>
> On 8/1/17 7:13 PM, DRC wrote:
>> Ugh.  Unfortunately I think you're running into the aforementioned GDM
>> bug, then.  Even more unfortunately, I just re-tested it with Fedora 26,
>> and it's still there, so I'm sure it hasn't been fixed upstream yet.
>> Does anyone fix bugs anymore?!  Maybe I'm expecting too much for GDM to
>> load its own farking init script.
>>
>>
>> On 8/1/17 1:07 PM, sta...@ms... wrote:
>>> Thank you for your quick reply,
>>> I can at least provide more information regarding some of the short
>>> testing you had said to do.
>>> When using the command: xauth merge /etc/opt/VirtualGL/vgl_xauth_key
>>> It returns back nothing.
>>> When using the command: xdpyinfo -display :0
>>> It returns: xdpyinfo:  unable to open display ":0".
>>> (also these commands are being done through ssh. If they need to be done
>>> directly I can)
>>>
>>> Further more When adding the line you requested into the
>>> /etc/mdm/Init/Default script (which is in the correct location) did not
>>> actually create file/folder that it should have. I wanted to verify that
>>> mdm is actually the display manager and it is (command used: cat
>>> /etc/X11/default-display-manager).
>>>
>>> I made sure that I was part of the vglusers group.
>>>
>>> I tried to redirect the output of the vglgenkey but it seems to not be
>>> touching the /etc/mdm/Init/Default script.
>>>
>>> I hope this sheds a little more light onto what is going on and if there
>>> is anything further I can provide i'd be happy to do as such. I
>>> appreciate your time suggestions.
>>> Thank you,
>>>
>>> Marshall Stageberg
>>> Michigan State University
>>>
>>> Quoting DRC via VirtualGL-Users <vir...@li...>:
>>>
>>>> The first thing to try, as indicated in the "Sanity Check" section of
>>>> the docs
>>>> (https://urldefense.proofpoint.com/v2/url?u=https-3A__cdn.rawgit.com_VirtualGL_virtualgl_master_doc_index.html-23hd006002&d=DwICAg&c=nE__W8dFE-shTxStwXtp0A&r=is60q22IcyABEXnsSrzYNQ&m=o2rLPFelD8-NEVKAzy69YMV_WeYzc4l-R3NRWKaDzLg&s=nlo4lHrLMFYP_Xf98qQ9cYZp7a4TF08aoFc0x7h2EEY&e=
>>>> ),
>>>> is:
>>>>
>>>>   xauth merge /etc/opt/VirtualGL/vgl_xauth_key
>>>>   xdpyinfo -display :0
>>>>
>>>> That may reveal the underlying cause of the problem.  If the xauth
>>>> command fails, then check the following:
>>>>
>>>> - Make sure that your installation of MDM keeps its initialization
>>>> script in the standard place (/etc/mdm/Init/Default).  I've tested Linux
>>>> Mint before, so it should work, but maybe they moved something in a more
>>>> recent release.
>>>>
>>>> - Make sure that /etc/mdm/Init/Default is being modified properly by
>>>> vglserver_config.  It should have a 'vglgenkey' line at the top if all
>>>> is well.
>>>>
>>>> - Edit /etc/mdm/Init/Default and add 'echo here >/tmp/here', then
>>>> restart the display manager and verify that /tmp/here is created.  If
>>>> not, then the display manager isn't executing the initialization script
>>>> as it should.  This is known to be an issue with some versions of GDM:
>>>> https://urldefense.proofpoint.com/v2/url?u=https-3A__bugzilla.redhat.com_show-5Fbug.cgi-3Fid-3D851769&d=DwICAg&c=nE__W8dFE-shTxStwXtp0A&r=is60q22IcyABEXnsSrzYNQ&m=o2rLPFelD8-NEVKAzy69YMV_WeYzc4l-R3NRWKaDzLg&s=Rn4PmhNvGhcUbs7hyAtSVAqsn3QY5YkZOj6UmgSE7nI&e=
>>>> .  Maybe the issue
>>>> creeped into MATE as well?  (I hope not.)
>>>>
>>>> - Make sure you added yourself to the vglusers group and logged
>>>> out/logged back in to activate the new group permissions.
>>>>
>>>> - Edit /etc/mdm/Init/Default and redirect the output of vglgenkey to a
>>>> file.  Restart the DM and look at the file to see if there are any error
>>>> messages reported by vglgenkey.
>>>>
>>>> This is generally the process I use when diagnosing failures to access
>>>> display :0.  Thus far, I have yet to encounter such a failure that
>>>> wasn't due to one of the issues above.
>>>>
>>>> On 8/1/17 9:54 AM, sta...@ms... wrote:
>>>>> Hi All,
>>>>> I have setup a headless server with 2 - Nvidia GTX 1080i video cards
>>>>> using Linux Mint 18.1 as the OS. I am using the Nvidia 384.59 drivers.
>>>>> Output from nvidia-xconfig --query-gpu-info:
>>>>> Number of GPUs: 2
>>>>>
>>>>> GPU #0:
>>>>>   Name      : GeForce GTX 1080 Ti
>>>>>   UUID      : GPU-cc7afa9d-a61d-950a-0f45-511bd6f1f340
>>>>>   PCI BusID : PCI:4:0:0
>>>>>
>>>>>   Number of Display Devices: 0
>>>>>
>>>>>
>>>>> GPU #1:
>>>>>   Name      : GeForce GTX 1080 Ti
>>>>>   UUID      : GPU-def710e8-b54f-d87d-be43-538d7395c0d8
>>>>>   PCI BusID : PCI:65:0:0
>>>>>
>>>>>   Number of Display Devices: 0
>>>>>
>>>>> Pertinent output from lspci:
>>>>> 04:00.0 VGA compatible controller: NVIDIA Corporation Device 1b06
>>>>> (rev a1)
>>>>> 41:00.0 VGA compatible controller: NVIDIA Corporation Device 1b06
>>>>> (rev a1)
>>>>>
>>>>> I am using the default mdm display manager (comes stock with linux mint)
>>>>> When I plug a monitor into the computer directly all video functions
>>>>> work correctly (glxgears, glxinfo, etc..)
>>>>>
>>>>> I setup the vncserver and then use the vncviewer to connect to the
>>>>> headless server. When I try to execute vglrun glxgears I get the error:
>>>>> [VGL] ERROR: Could not open display :0.
>>>>> I have ran the vglserver_config script with stopping the following
>>>>> modules: nvidia_drm, nvidia_uvm, nvidia_modeset, and nvidia. I also
>>>>> stopped the killed the nvidia persistence process prior to the setup
>>>>> along with mdm. The config seemed to complete perfectly and I have
>>>>> restarted the computer.
>>>>>
>>>>> I appreciate any further advice with this issue.
>>>>> Thank you,
>>>>>
>>>>> Marshall Stageberg
>>>>> Michigan State University
>
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org!  
> https://urldefense.proofpoint.com/v2/url?u=http-3A__sdm.link_slashdot&d=DwICAg&c=nE__W8dFE-shTxStwXtp0A&r=is60q22IcyABEXnsSrzYNQ&m=SyCf0kY2YDi4Tez8dF29jMd-L8WEuwnKYvK6RuTVVmM&s=F1Wc8JBKjtnqHWJ9foKn3gT2yA-TwIFZi846kVxUDNc&e=
> _______________________________________________
> VirtualGL-Users mailing list
> Vir...@li...
> https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.sourceforge.net_lists_listinfo_virtualgl-2Dusers&d=DwICAg&c=nE__W8dFE-shTxStwXtp0A&r=is60q22IcyABEXnsSrzYNQ&m=SyCf0kY2YDi4Tez8dF29jMd-L8WEuwnKYvK6RuTVVmM&s=Kj7Lo51y2q3tBSMGseAG7ih90I8febAcl6uxh9XpviY&e=
>