fix(ci): add imagePullPolicy: IfNotPresent to training workflow

Kubernetes defaults to Always for :latest tags, forcing registry
round-trips that fail on fresh GPU nodes where containerd HTTP-only
registry config has a race condition with HTTPS fallback.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
jgrusewski
2026-03-06 12:49:34 +01:00
parent 2f8fa1ab19
commit 4e0d1fcbe6

View File

@@ -26,7 +26,7 @@ spec:
parameters:
- name: model
- name: gpu-pool
value: ci-training-h100
value: ci-training-h100 # Pools: ci-training-h100, ci-training-l40s, ci-training-h100x2
- name: hyperopt-trials
value: "20"
- name: hyperopt-epochs
@@ -97,6 +97,7 @@ spec:
effect: NoSchedule
container:
image: gitlab-registry.foxhunt.svc.cluster.local:5000/root/foxhunt/foxhunt-training-runtime:latest
imagePullPolicy: IfNotPresent
command: ["/bin/sh", "-c"]
args:
- |
@@ -118,6 +119,7 @@ spec:
k8s.scaleway.com/pool-name: platform
container:
image: gitlab-registry.foxhunt.svc.cluster.local:5000/root/foxhunt/foxhunt-runtime:latest
imagePullPolicy: IfNotPresent
command: ["/bin/sh", "-c"]
args:
- |
@@ -168,6 +170,7 @@ spec:
effect: NoSchedule
container:
image: gitlab-registry.foxhunt.svc.cluster.local:5000/root/foxhunt/foxhunt-training-runtime:latest
imagePullPolicy: IfNotPresent
command: ["/bin/sh", "-c"]
args:
- |
@@ -264,6 +267,7 @@ spec:
effect: NoSchedule
container:
image: gitlab-registry.foxhunt.svc.cluster.local:5000/root/foxhunt/foxhunt-training-runtime:latest
imagePullPolicy: IfNotPresent
command: ["/bin/sh", "-c"]
args:
- |
@@ -355,6 +359,7 @@ spec:
effect: NoSchedule
container:
image: gitlab-registry.foxhunt.svc.cluster.local:5000/root/foxhunt/foxhunt-training-runtime:latest
imagePullPolicy: IfNotPresent
command: ["/bin/sh", "-c"]
args:
- |
@@ -429,6 +434,7 @@ spec:
k8s.scaleway.com/pool-name: platform
container:
image: gitlab-registry.foxhunt.svc.cluster.local:5000/root/foxhunt/foxhunt-runtime:latest
imagePullPolicy: IfNotPresent
command: ["/bin/sh", "-c"]
args:
- |