Questions to Determine¶
- Which DaemonSet pods are present on the node being drained?
- Is the
--ignore-daemonsetsflag appropriate for this scenario? - Are any of the DaemonSet pods critical for node-level cleanup before shutdown?
- Do any DaemonSet pods have local data that needs to be flushed before eviction?
- Are there non-DaemonSet pods on the node that also need to be evicted?
- Will the Fluentd DaemonSet lose buffered log data if the pod is terminated?
- Is there a PodDisruptionBudget on any of the DaemonSet pods?
- What is the graceful shutdown behavior of each DaemonSet pod?