H2O's Driverless AI .deb package failing to install

221 Views Asked by At

I am running Ubuntu 18.04 with CUDA 9.1. I have downloaded the CUDA 9.0 Driverless AI .deb installation from the H2O website's download page: https://www.h2o.ai/driverless-ai-download/

Unfortunately I am getting the following:

kevin@Ubuntu-XPS:~/Downloads$ sudo dpkg -i dai_1.1.4_amd64.deb 
[sudo] password for kevin: 
Selecting previously unselected package dai.
(Reading database ... 486543 files and directories currently installed.)
Preparing to unpack dai_1.1.4_amd64.deb ...
Unpacking dai (1.1.4) ...
Setting up dai (1.1.4) ...
User configuration file /etc/dai/User.conf already exists.
Group configuration file /etc/dai/Group.conf already exists.
Configured user in /etc/dai/User.conf is 'dai'.
Configured group in /etc/dai/Group.conf is 'dai'.
Group 'dai' already exists.
User 'dai' already exists.
Creating /opt/h2oai/dai/tmp...
Creating /opt/h2oai/dai/home...
Creating /opt/h2oai/dai/log... (Note for systemd users this log dir will be unused; use journalctl instead.)
Adding systemd configuration files in /etc/systemd/system...
Failed to lookup unit file state: Invalid argument
dpkg: error processing package dai (--install):
 installed dai package post-installation script subprocess returned error exit status 62
Errors were encountered while processing:
 dai

Could it be any of the following problems:

  • Driverless AI is not yet supported on Ubuntu 18.04?
  • Driverless AI is not yet supported on CUDA 9.1?
  • The .deb file ends in amd64 but I have an Intel i7? (Although the Ubuntu 18.04 image file from the official downloads page also ends in amd64)
1

There are 1 best solutions below

2
On

As of Driverless AI 1.2.2 the only version of Ubuntu that is tested is Ubuntu 16.04.

If you are having dpkg installation problems, you can try the (beta) “TAR SH” installation package style of the most recent version on h2o.ai/download (today that is 1.2.2).

(This additional new approach just unpacks into the current directory.)

Additionally, CUDA 9.1 is not supported (and probably never will be, since CUDA 9.2 is already out). But that’s not the cause of the error above.