John Lockman

lock down mariadb

1

**Is your feature request related to a problem? Please describe.** please lock down access to mariadb. **Describe the solution you'd like** restrict access to mariadb for root login only

Slurm Settings for Large Clusters

4

**Is your feature request related to a problem? Please describe.** There are a few settings that need to be updated when running large scale Slurm clusters described here: https://slurm.schedmd.com/big_sys.html **Describe...

use Nvidia galaxy role for driver install

8

**Is your feature request related to a problem? Please describe.** Nvidia has a galaxy collection for the nvidia driver that we should investigate: https://galaxy.ansible.com/nvidia/nvidia_driver **Describe the solution you'd like** Evaluate...

Upgrade nvidia-device-plugin deployment

1

please upgrade nvidia-device-plugin to v0.12.2 to support Cuda 11 **Describe the solution you'd like** [Nvidia k8s device plugin](https://github.com/NVIDIA/k8s-device-plugin#deploying-with-gpu-feature-discovery-for-automatic-node-labels) Deploying with gpu-feature-discovery for automatic node labels As of v0.12.0, the device...

volcano should be an optional install

3

**Describe the bug** the volcano.sh k8s scheduler is currently installed by default and should only be installed when requested.

PMI2 interface support?

1

I understand PMI2 interface is "mostly" supported, is there any documentation on what is supported from the standard?

Configure ib0 for overlay/virtual netdevice

3

Can you elaborate on the steps necessary to "...configure ib0 or appropriate IPoIB netdevice as the parent netdevice for creating overlay/virtual netdevices."? Is this supposed to work if you have...

**Is your feature request related to a problem? Please describe.** versions need updating, let's start with v1.22.14-0 **Describe the solution you'd like** keep older versions, just add updated version.

**Is your feature request related to a problem? Please describe.** The [destroy ](https://github.com/dellhpc/omnia/blob/devel/destroy.yml)role could be improved, it currently only tears down a k8s cluster. It should also: * tear down...

Update xilinx fpga device plugin location

1

**Describe the bug** in https://github.com/dellhpc/omnia/blob/devel/roles/k8s_start_services/vars/main.yml update line 80: `fpga_device_plugin_yaml_url: https://raw.githubusercontent.com/Xilinx/FPGA_as_a_Service/master/k8s-fpga-device-plugin/fpga-device-plugin.yml` with updated URL: `fpga_device_plugin_yaml_url: https://raw.githubusercontent.com/Xilinx/FPGA_as_a_Service/master/k8s-device-plugin/k8s-device-plugin.yml`