Skip to content
Open
Changes from all commits
Commits
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
38 changes: 36 additions & 2 deletions cmd/installer/goods/support/host-support-bundle.tmpl.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -197,7 +197,7 @@ spec:
- run:
collectorName: curl-replicated-app
command: sh
args:
args:
- -c
- |
if [ -n "{{ .HTTPSProxy }}" ]; then
Expand Down Expand Up @@ -398,7 +398,6 @@ spec:
- -c
- |
pat='(clamav|sophos|esets_daemon|fsav|symantec|mfend|ds_agent|kav|bdagent|s1agent|falcon|illumio|xagt|wdavdaemon|mdatp)'

if command -v pgrep >/dev/null 2>&1; then
pgrep -afi "$pat"
else
Expand Down Expand Up @@ -733,6 +732,41 @@ spec:
- pass:
when: "false"
message: "No signs of hostname changes found"
- textAnalyze:
checkName: PodDisruptionBudget blocking upgrade
fileName: host-collectors/run-host/k0scontroller-logs.txt
regex: '(error when evicting pods|Cannot evict pod as it would violate the pod''s disruption budget)'
outcomes:
- fail:
when: "true"
message: |-
PodDisruptionBudgets are preventing pod eviction during cluster upgrade.

The k0scontroller cannot evict pods, causing the autopilot upgrade process to hang.
This makes the cluster unable to complete upgrades.

To resolve this issue:

1. List all PodDisruptionBudgets in the cluster:
kubectl get pdb -A

2. Identify which PDBs are blocking eviction. Look for PDBs with:
- minAvailable set too high (not allowing enough pods to be evicted)
- maxUnavailable set too low (too restrictive)

3. Temporarily adjust or remove the problematic PDBs:
kubectl delete pdb <pdb-name> -n <namespace>

Or edit to allow more disruption:
kubectl edit pdb <pdb-name> -n <namespace>

4. After the upgrade completes, restore the PDBs if needed.

WARNING: Modifying PDBs reduces availability guarantees. Coordinate with
application owners and consider scheduling during maintenance windows.
- pass:
when: "false"
message: No PodDisruptionBudget eviction failures detected
- textAnalyze:
checkName: Check if localhost resolves to 127.0.0.1
fileName: host-collectors/run-host/localhost-ips.txt
Expand Down
Loading