NVIDIA / gpu-operator / 34854940802

44%

main: 37%

push

github

use a direct reader to check for resource claim for DRA

Signed-off-by: Karthikeyan Valliyurnatt <kvalliyurnat@nvidia.com>

Add configurable securityContext and dnsConfig for the operator Deployment

Default the gpu-operator Deployment and Helm hook Jobs to a restricted
pod securityContext (runAsNonRoot, RuntimeDefault seccomp), configurable
via operator.securityContext. The container securityContext (no
privilege escalation, read-only root, drop ALL) is fixed in the
templates rather than exposed as a values key, to limit API surface.
Do not pin UID/GID so OpenShift MustRunAsRange can allocate; the image
USER 1000:1000 still applies on vanilla Kubernetes. Expose discoverable
dnsPolicy and dnsConfig values keys (dnsPolicy empty keeps ClusterFirst).
Operand DaemonSets are unchanged; they still need host privileges.

Fixes #2533

Signed-off-by: Dundy Pasupuleti <pasupuletidundy@gmail.com>

23 of 25 new or added lines in 1 file covered. (92.0%)

5746 of 12973 relevant lines covered (44.29%)

0.44 hits per line