Pods stuck Pending after node failure: why didn't Kubernetes reschedule them to the other node?Pods stuck Pending after node failure: a required NotIn node affinity and two topology spread defaults on a 2-node k3s cluster. How to check yours.24 min read·kkhdevs.hashnode.dev6h ago
Kubernetes CronJob pg_dump backup: does the dump end up on the same node as the database?Kubernetes CronJob pg_dump backup on k3s local-path: 5 of 7 dumps and every etcd snapshot were on the database's own disk. How to check yours.21 min read·kkhdevs.hashnode.dev15h ago
Sixteen pods evicted for disk, thirteen of them above the thresholdSixteen pods evicted on one k3s node in three waves. Thirteen above the 5 % threshold: evictionMinimumReclaim was 10 % and nothing requested ephemeral-storage.13 min read·kkhdevs.hashnode.dev1d ago
A pod that ran for six months crash-looped on its first restartcoturn:latest with imagePullPolicy Always: one restart in late September pulled 4.18.0, which had removed --no-dtls. 280 restarts a day for 45 days, alert unread.11 min read·kkhdevs.hashnode.dev1d ago
ImagePullBackOff on an image nobody had touched since MarchA registry GC with --delete-untagged removed OCI index children in March; the node cache hid it until a disk-pressure eviction in November.10 min read·kkhdevs.hashnode.dev2d ago
