DGX | unmet dependencies error

While trying to install the below list of packages, unmet dependencies error

Step 1

sudo apt-get install protobuf-compiler python-pil python-lxml python-tk

error




sudo apt-get install protobuf-compiler python-pil python-lxml python-tk
Reading package lists... Done


Building dependency tree       
Reading state information... Done
protobuf-compiler is already the newest version (3.0.0-9.1ubuntu1).
You might want to run 'apt --fix-broken install' to correct these.
The following packages have unmet dependencies:
 nvidia-cuda-toolkit : Depends: nvidia-cuda-dev (= 9.1.85-3ubuntu1) but it is not going to be installed
E: Unmet dependencies. Try 'apt --fix-broken install' with no packages (or specify a solution).

Step 2
Tried

sudo apt --fix-broken install

Error during Step 2

sudo apt --fix-broken install
Reading package lists... Done
Building dependency tree       
Reading state information... Done
Correcting dependencies... Done
The following additional packages will be installed:
  nvidia-cuda-dev
Recommended packages:
  libvdpau-dev libnvcuvid1
The following NEW packages will be installed:
  nvidia-cuda-dev
0 upgraded, 1 newly installed, 0 to remove and 534 not upgraded.
1 not fully installed or removed.
Need to get 0 B/263 MB of archives.
After this operation, 734 MB of additional disk space will be used.
Do you want to continue? [Y/n] Y
(Reading database ... 228468 files and directories currently installed.)
Preparing to unpack .../nvidia-cuda-dev_9.1.85-3ubuntu1_amd64.deb ...
Unpacking nvidia-cuda-dev (9.1.85-3ubuntu1) ...
dpkg: error processing archive /var/cache/apt/archives/nvidia-cuda-dev_9.1.85-3ubuntu1_amd64.deb (--unpack):
 trying to overwrite '/usr/include/nvml.h', which is also in package datacenter-gpu-manager 1:1.5.9
dpkg-deb: error: paste subprocess was killed by signal (Broken pipe)
Errors were encountered while processing:
 /var/cache/apt/archives/nvidia-cuda-dev_9.1.85-3ubuntu1_amd64.deb
E: Sub-process /usr/bin/dpkg returned an error code (1)

Step 3
Tried

sudo dpkg --configure -a

Error

dpkg: dependency problems prevent configuration of nvidia-cuda-toolkit:
 nvidia-cuda-toolkit depends on nvidia-cuda-dev (= 9.1.85-3ubuntu1); however:
  Package nvidia-cuda-dev is not installed.

dpkg: error processing package nvidia-cuda-toolkit (--configure):
 dependency problems - leaving unconfigured
Errors were encountered while processing:
 nvidia-cuda-toolkit

Status of NVIDIA-SMI

nvidia-smi
Sun Sep 29 18:51:56 2019       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 410.104      Driver Version: 410.104      CUDA Version: 10.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Tesla V100-DGXS...  On   | 00000000:07:00.0  On |                    0 |
| N/A   36C    P0    40W / 300W |    426MiB / 32475MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   1  Tesla V100-DGXS...  On   | 00000000:08:00.0 Off |                    0 |
| N/A   35C    P0    40W / 300W |      0MiB / 32478MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   2  Tesla V100-DGXS...  On   | 00000000:0E:00.0 Off |                    0 |
| N/A   35C    P0    38W / 300W |      0MiB / 32478MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   3  Tesla V100-DGXS...  On   | 00000000:0F:00.0 Off |                    0 |
| N/A   35C    P0    39W / 300W |      0MiB / 32478MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |

It looks like it’s trying to pull in an old version of CUDA stuff - the DGX release you’re on (judging by the driver version) is CUDA 10 based, so pulling in CUDA 9.1 stuff isn’t going to be pretty.

I’d recommend using the NGC CUDA container as your starting point, instead of DGX OS. (e.g., https://ngc.nvidia.com/catalog/containers/nvidia:cuda/tags )

If you really need this to work on DGX OS as-is, our NVIDIA Enterprise Support team would be happy to help! You can create a new ticket by sending an email to EnterpriseSupport@nvidia.com , or through the NVIDIA Enterprise Support Portal. (See https://www.nvidia.com/en-us/support/enterprise/ )