Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

update: gpu config #103

Merged
merged 4 commits into from
Oct 3, 2024
Merged

update: gpu config #103

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
e7f1d09
update: remove default taint
codekow Oct 3, 2024
22e90ff
update: carve out gpu taint
codekow Oct 3, 2024
a06ee59
update: scripts
codekow Oct 3, 2024
82272a9
cleanup
codekow Oct 3, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 0 additions & 6 deletions components/app-configs/nvidia-gpu-verification/overlays/toleration/kustomization.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,9 +16,6 @@ patches:
- effect: NoSchedule
key: nvidia.com/gpu
operator: Exists
- effect: NoSchedule
key: nvidia-gpu-only
operator: Exists
- target:
version: v1
kind: Pod
Expand All @@ -29,6 +26,3 @@ patches:
- effect: NoSchedule
key: nvidia.com/gpu
operator: Exists
- effect: NoSchedule
key: nvidia-gpu-only
operator: Exists
3 changes: 0 additions & 3 deletions components/operators/devspaces/instance/overlays/nvidia-gpu/kustomization.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,6 +28,3 @@ patches:
- effect: NoSchedule
key: nvidia.com/gpu
operator: Exists
- effect: NoSchedule
key: nvidia-gpu-only
operator: Exists
3 changes: 0 additions & 3 deletions components/operators/gpu-operator-certified/instance/base/cluster-policy.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,9 +55,6 @@ spec:
- effect: NoSchedule
key: nvidia.com/gpu
operator: Exists
- effect: NoSchedule
key: nvidia-gpu-only
operator: Exists
devicePlugin:
enabled: true
config:
Expand Down
1 change: 1 addition & 0 deletions components/operators/gpu-operator-certified/instance/components/aws-gpu-machineset/job.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,3 +8,4 @@ INSTANCE_TYPE=${INSTANCE_TYPE:-g4dn.4xlarge}
ocp_aws_cluster || exit 0
ocp_aws_create_gpu_machineset "${INSTANCE_TYPE}"
ocp_create_machineset_autoscale
# ocp_aws_taint_gpu_machineset
29 changes: 17 additions & 12 deletions components/operators/gpu-operator-certified/instance/components/aws-gpu-machineset/ocp.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,9 +6,11 @@
# ocp_aws_cluster
# ocp_aws_create_gpu_machineset
# ocp_aws_clone_worker_machineset
# ocp_create_machineset_autoscale'
# ocp_aws_taint_gpu_machineset
# ocp_create_machineset_autoscale
# '

# for function in $FUNCTIONS
# for function in ${FUNCTIONS}
# do
# extract_function $function scripts/library/ocp.sh >> tmp
# echo >> tmp
Expand Down Expand Up @@ -45,16 +47,7 @@ ocp_aws_create_gpu_machineset(){
oc -n openshift-machine-api \
patch "${MACHINE_SET_TYPE}" \
--type=merge --patch '{"spec":{"template":{"spec":{"metadata":{"labels":{"node-role.kubernetes.io/gpu":""}}}}}}'

# taint nodes for gpu-only workloads
oc -n openshift-machine-api \
patch "${MACHINE_SET_TYPE}" \
--type=merge --patch '{"spec":{"template":{"spec":{"taints":[{"key":"nvidia.com/gpu","value":"","effect":"NoSchedule"}]}}}}'

# oc -n openshift-machine-api \
# patch "${MACHINE_SET_TYPE}" \
# --type=merge --patch '{"spec":{"template":{"spec":{"taints":[{"key":"nvidia-gpu-only","value":"","effect":"NoSchedule"}]}}}}'


# should use the default profile
# oc -n openshift-machine-api \
# patch "${MACHINE_SET_TYPE}" \
Expand Down Expand Up @@ -112,6 +105,18 @@ ocp_aws_clone_worker_machineset(){
--type=merge --patch '{"spec":{"template":{"spec":{"metadata":{"labels":{"node-role.kubernetes.io/'"${SHORT_NAME}"'":""}}}}}}'
}

ocp_aws_taint_gpu_machineset(){
INSTANCE_TYPE=${1:-g4dn.4xlarge}
MACHINE_SET_TYPE=$(oc -n openshift-machine-api get machinesets.machine.openshift.io -o name | grep "${INSTANCE_TYPE%.*}" | head -n1)

echo "Patching: ${MACHINE_SET_TYPE}"

# taint nodes for gpu-only workloads
oc -n openshift-machine-api \
patch "${MACHINE_SET_TYPE}" \
--type=merge --patch '{"spec":{"template":{"spec":{"taints":[{"key":"nvidia.com/gpu","value":"","effect":"NoSchedule"}]}}}}'
}

ocp_create_machineset_autoscale(){
MACHINE_MIN=${1:-0}
MACHINE_MAX=${2:-4}
Expand Down
5 changes: 0 additions & 5 deletions .../gpu-operator-certified/instance/components/time-sliced-4/patch-device-plugin-config.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,11 +5,6 @@ metadata:
data:
no-time-sliced: |-
version: v1
sharing:
timeSlicing:
resources:
- name: nvidia.com/gpu
replicas: 0
time-sliced: |-
version: v1
sharing:
Expand Down
5 changes: 0 additions & 5 deletions ...rs/gpu-operator-certified/instance/components/time-sliced/patch-device-plugin-config.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,11 +5,6 @@ metadata:
data:
no-time-sliced: |-
version: v1
sharing:
timeSlicing:
resources:
- name: nvidia.com/gpu
replicas: 0
time-sliced: |-
version: v1
sharing:
Expand Down
3 changes: 0 additions & 3 deletions ...ors/rhods-operator/instance/components/nvidia-gpu-accelerator-profile/nvidia-profile.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,3 @@ spec:
- effect: NoSchedule
key: nvidia.com/gpu
operator: Exists
- effect: NoSchedule
key: nvidia-gpu-only
operator: Exists
29 changes: 16 additions & 13 deletions scripts/library/ocp.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -243,15 +243,6 @@ ocp_aws_create_gpu_machineset(){
oc -n openshift-machine-api \
patch "${MACHINE_SET_TYPE}" \
--type=merge --patch '{"spec":{"template":{"spec":{"metadata":{"labels":{"node-role.kubernetes.io/gpu":""}}}}}}'

# taint nodes for gpu-only workloads
oc -n openshift-machine-api \
patch "${MACHINE_SET_TYPE}" \
--type=merge --patch '{"spec":{"template":{"spec":{"taints":[{"key":"nvidia.com/gpu","value":"","effect":"NoSchedule"}]}}}}'

# oc -n openshift-machine-api \
# patch "${MACHINE_SET_TYPE}" \
# --type=merge --patch '{"spec":{"template":{"spec":{"taints":[{"key":"nvidia-gpu-only","value":"","effect":"NoSchedule"}]}}}}'

# should use the default profile
# oc -n openshift-machine-api \
Expand All @@ -272,6 +263,18 @@ ocp_aws_create_gpu_machineset(){
--type=merge --patch '{"spec":{"template":{"spec":{"providerSpec":{"value":{"instanceType":"'"${INSTANCE_TYPE}"'"}}}}}}'
}

ocp_aws_taint_gpu_machineset(){
INSTANCE_TYPE=${1:-g4dn.4xlarge}
MACHINE_SET_TYPE=$(oc -n openshift-machine-api get machinesets.machine.openshift.io -o name | grep "${INSTANCE_TYPE%.*}" | head -n1)

echo "Patching: ${MACHINE_SET_TYPE}"

# taint nodes for gpu-only workloads
oc -n openshift-machine-api \
patch "${MACHINE_SET_TYPE}" \
--type=merge --patch '{"spec":{"template":{"spec":{"taints":[{"key":"nvidia.com/gpu","value":"","effect":"NoSchedule"}]}}}}'
}

ocp_create_machineset_autoscale(){
MACHINE_MIN=${1:-0}
MACHINE_MAX=${2:-4}
Expand Down Expand Up @@ -332,8 +335,8 @@ ocp_set_scheduler_profile(){
SCHED_PROFILE=${1:-LowNodeUtilization}

# LowNodeUtilization, HighNodeUtilization, NoScoring
echo "see https://docs.openshift.com/container-platform/4.11/nodes/scheduling/nodes-scheduler-profiles.html"
echo "OPTIONS: LowNodeUtilization, HighNodeUtilization, NoScoring"
echo "see https://docs.openshift.com/container-platform/4.16/nodes/scheduling/nodes-scheduler-profiles.html"
echo "OPTIONS: LowNodeUtilization (default), HighNodeUtilization, NoScoring"
echo "SCHED_PROFILE: ${SCHED_PROFILE}"

oc patch schedulers.config.openshift.io/cluster --type merge --patch '{"spec":{"profile": "'"${SCHED_PROFILE}"'"}}'
Expand Down Expand Up @@ -411,13 +414,13 @@ ocp_ack_upgrade_4.13(){
}

ocp_gpu_taint_nodes(){
oc adm taint node -l node-role.kubernetes.io/gpu nvidia-gpu-only=:NoSchedule --overwrite
oc adm taint node -l node-role.kubernetes.io/gpu nvidia.com/gpu=:NoSchedule --overwrite
oc adm drain -l node-role.kubernetes.io/gpu --ignore-daemonsets --delete-emptydir-data
oc adm uncordon -l node-role.kubernetes.io/gpu
}

ocp_gpu_untaint_nodes(){
oc adm taint node -l node-role.kubernetes.io/gpu nvidia-gpu-only=:NoSchedule-
oc adm taint node -l node-role.kubernetes.io/gpu nvidia.com/gpu=:NoSchedule-
}

ocp_gpu_label_nodes_from_nfd(){
Expand Down
3 changes: 0 additions & 3 deletions workshop/wip/parasol-insurance/01-setup/kustomization.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -33,9 +33,6 @@ patches:
- effect: NoSchedule
key: nvidia.com/gpu
operator: Exists
- effect: NoSchedule
key: nvidia-gpu-only
operator: Exists
- path: exclude-sc-ceph-rbd.yaml
- path: exclude-sc-gp3.yaml
- path: exclude-cm.yaml
Expand Down
Loading