actions
/
unity-builder
mirror of https://github.com/game-ci/unity-builder.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

pr feedback - fix taint removal syntax

pull/767/head
Frostebite 2025-12-29 18:26:09 +00:00
parent 355551c72e
commit 45e7ed0fcb
1 changed files with 38 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

40
.github/workflows/cloud-runner-integrity.yml vendored
View File

@ -238,13 +238,20 @@ jobs:
fi fi
done done
# If disk pressure taint is still present after cleanup, manually remove it (CI only) # If disk pressure taint is still present after cleanup, manually remove it (CI only)
# Try multiple times as Kubernetes may re-add it if condition persists
if kubectl describe nodes 2>/dev/null | grep -q "node.kubernetes.io/disk-pressure"; then if kubectl describe nodes 2>/dev/null | grep -q "node.kubernetes.io/disk-pressure"; then
echo "WARNING: Disk pressure taint still present after cleanup. Manually removing taint for CI..." echo "WARNING: Disk pressure taint still present after cleanup. Manually removing taint for CI..."
NODE_NAMES=$(kubectl get nodes -o name 2>/dev/null | sed 's/node\///' || echo "") NODE_NAMES=$(kubectl get nodes -o name 2>/dev/null | sed 's/node\///' || echo "")
for node in $NODE_NAMES; do for node in $NODE_NAMES; do
# Try removing with NoSchedule effect (most common)
kubectl taint nodes "$node" node.kubernetes.io/disk-pressure:NoSchedule- 2>/dev/null || true
# Also try without effect specifier
kubectl taint nodes "$node" node.kubernetes.io/disk-pressure- 2>/dev/null || true kubectl taint nodes "$node" node.kubernetes.io/disk-pressure- 2>/dev/null || true
# Use patch as fallback
kubectl patch node "$node" -p '{"spec":{"taints":[]}}' 2>/dev/null || true
done done
echo "Taint removed. Checking nodes..." sleep 2
echo "Taint removal attempted. Checking nodes..."
kubectl describe nodes | grep -i taint || echo "No taints found" kubectl describe nodes | grep -i taint || echo "No taints found"
fi fi
# Wait for disk pressure condition to clear (not just taint) # Wait for disk pressure condition to clear (not just taint)
@ -264,9 +271,20 @@ jobs:
echo "WARNING: Disk pressure condition still exists. Removing taint and waiting 10 seconds..." echo "WARNING: Disk pressure condition still exists. Removing taint and waiting 10 seconds..."
NODE_NAMES=$(kubectl get nodes -o name 2>/dev/null | sed 's/node\///' || echo "") NODE_NAMES=$(kubectl get nodes -o name 2>/dev/null | sed 's/node\///' || echo "")
for node in $NODE_NAMES; do for node in $NODE_NAMES; do
# Try removing with NoSchedule effect (most common)
kubectl taint nodes "$node" node.kubernetes.io/disk-pressure:NoSchedule- 2>/dev/null || true
# Also try without effect specifier
kubectl taint nodes "$node" node.kubernetes.io/disk-pressure- 2>/dev/null || true kubectl taint nodes "$node" node.kubernetes.io/disk-pressure- 2>/dev/null || true
# Use patch as fallback to remove all taints
kubectl patch node "$node" -p '{"spec":{"taints":[]}}' 2>/dev/null || true
done done
sleep 10 sleep 10
# Verify taint is actually removed
if kubectl describe nodes 2>/dev/null | grep -q "node.kubernetes.io/disk-pressure"; then
echo "ERROR: Taint still present after removal attempts. This may cause pod scheduling issues."
else
echo "Taint successfully removed."
fi
fi fi
- name: Run cloud-runner-image test (validate image creation) - name: Run cloud-runner-image test (validate image creation)
timeout-minutes: 10 timeout-minutes: 10
@ -488,13 +506,20 @@ jobs:
fi fi
done done
# If disk pressure taint is still present after cleanup, manually remove it (CI only) # If disk pressure taint is still present after cleanup, manually remove it (CI only)
# Try multiple times as Kubernetes may re-add it if condition persists
if kubectl describe nodes 2>/dev/null | grep -q "node.kubernetes.io/disk-pressure"; then if kubectl describe nodes 2>/dev/null | grep -q "node.kubernetes.io/disk-pressure"; then
echo "WARNING: Disk pressure taint still present after cleanup. Manually removing taint for CI..." echo "WARNING: Disk pressure taint still present after cleanup. Manually removing taint for CI..."
NODE_NAMES=$(kubectl get nodes -o name 2>/dev/null | sed 's/node\///' || echo "") NODE_NAMES=$(kubectl get nodes -o name 2>/dev/null | sed 's/node\///' || echo "")
for node in $NODE_NAMES; do for node in $NODE_NAMES; do
# Try removing with NoSchedule effect (most common)
kubectl taint nodes "$node" node.kubernetes.io/disk-pressure:NoSchedule- 2>/dev/null || true
# Also try without effect specifier
kubectl taint nodes "$node" node.kubernetes.io/disk-pressure- 2>/dev/null || true kubectl taint nodes "$node" node.kubernetes.io/disk-pressure- 2>/dev/null || true
# Use patch as fallback
kubectl patch node "$node" -p '{"spec":{"taints":[]}}' 2>/dev/null || true
done done
echo "Taint removed. Checking nodes..." sleep 2
echo "Taint removal attempted. Checking nodes..."
kubectl describe nodes | grep -i taint || echo "No taints found" kubectl describe nodes | grep -i taint || echo "No taints found"
fi fi
# Wait for disk pressure condition to clear (not just taint) # Wait for disk pressure condition to clear (not just taint)
@ -514,9 +539,20 @@ jobs:
echo "WARNING: Disk pressure condition still exists. Removing taint and waiting 10 seconds..." echo "WARNING: Disk pressure condition still exists. Removing taint and waiting 10 seconds..."
NODE_NAMES=$(kubectl get nodes -o name 2>/dev/null | sed 's/node\///' || echo "") NODE_NAMES=$(kubectl get nodes -o name 2>/dev/null | sed 's/node\///' || echo "")
for node in $NODE_NAMES; do for node in $NODE_NAMES; do
# Try removing with NoSchedule effect (most common)
kubectl taint nodes "$node" node.kubernetes.io/disk-pressure:NoSchedule- 2>/dev/null || true
# Also try without effect specifier
kubectl taint nodes "$node" node.kubernetes.io/disk-pressure- 2>/dev/null || true kubectl taint nodes "$node" node.kubernetes.io/disk-pressure- 2>/dev/null || true
# Use patch as fallback to remove all taints
kubectl patch node "$node" -p '{"spec":{"taints":[]}}' 2>/dev/null || true
done done
sleep 10 sleep 10
# Verify taint is actually removed
if kubectl describe nodes 2>/dev/null | grep -q "node.kubernetes.io/disk-pressure"; then
echo "ERROR: Taint still present after removal attempts. This may cause pod scheduling issues."
else
echo "Taint successfully removed."
fi
fi fi
- name: Run cloud-runner-s3-steps test (validate S3 operations with K8s) - name: Run cloud-runner-s3-steps test (validate S3 operations with K8s)
timeout-minutes: 30 timeout-minutes: 30