Dunia K9: Pentadbir Kluster

Sandaran dan pemulihan etcd

Anda akan belajar: fotokopi buku pesanan dengan etcdctl, dan pulihkan objek yang hilang dengan etcdutl.

Teks BM belum disemak oleh penutur asli.

Disemak dengan Kubernetes 1.37 pada 1 Nov 2026

Masalahnya

Pukul 2 pagi, seseorang padam Deployment shop. YAML-nya tak pernah disimpan dalam Git. Apa yang masih boleh pulihkannya?

Tengok

Anda di sini: kluster kind-dev namespace default

~/shop $ ssh dev-control-planeroot@dev-control-plane:~# etcdctl --endpoints=https://127.0.0.1:2379 --cacert=/etc/kubernetes/pki/etcd/ca.crt --cert=/etc/kubernetes/pki/etcd/server.crt --key=/etc/kubernetes/pki/etcd/server.key snapshot save /opt/backup.db{"level":"info","ts":"2026-10-31T01:00:22.252272Z","caller":"snapshot/v3_snapshot.go:83","msg":"created temporary db file","path":"/opt/backup.db.part"}{"level":"info","ts":"2026-10-31T01:00:22.255184Z","logger":"client","caller":"v3/maintenance.go:236","msg":"opened snapshot stream; downloading"}{"level":"info","ts":"2026-10-31T01:00:22.258272Z","caller":"snapshot/v3_snapshot.go:96","msg":"fetching snapshot","endpoint":"https://127.0.0.1:2379"}{"level":"info","ts":"2026-10-31T01:00:22.260284Z","logger":"client","caller":"v3/maintenance.go:302","msg":"completed snapshot read; closing"}{"level":"info","ts":"2026-10-31T01:00:22.263372Z","caller":"snapshot/v3_snapshot.go:111","msg":"fetched snapshot","endpoint":"https://127.0.0.1:2379","size":"2.4 MB","took":"9.935898ms","etcd-version":"3.7.0"}{"level":"info","ts":"2026-10-31T01:00:22.266284Z","caller":"snapshot/v3_snapshot.go:121","msg":"saved","path":"/opt/backup.db"}Snapshot saved at /opt/backup.dbServer version 3.7.0root@dev-control-plane:~# exitlogoutConnection to dev-control-plane closed.

ialah fotokopi buku pesanan. etcdctl minta salinan itu daripada etcd yang sedang berjalan. Simpan salinan itu di luar kluster.

Cuba

Bina arahan simpan. Setiap kepingan tunjuk maksudnya.

Ketik kepingan ikut turutan yang betul.

…

Tengok

Anda di sini: kluster kind-dev namespace default

~/shop $ ssh dev-control-planeroot@dev-control-plane:~# etcdctl snapshot save /opt/backup.db{"level":"info","ts":"2026-10-31T01:00:22.252272Z","caller":"snapshot/v3_snapshot.go:83","msg":"created temporary db file","path":"/opt/backup.db.part"}^C# etcdctl is still waiting here: it never got in. You pressed Ctrl+C.root@dev-control-plane:~# exitlogoutConnection to dev-control-plane closed.

Perangkap sebenar: tanpa flag sijil, etcdctl cetak satu baris dan tunggu selamanya. Tekan Ctrl+C. Kemudian tambah --cacert, --cert dan --key.

Cuba

Giliran anda. Buka node pejabat, simpan snapshot, kemudian semak.

Anda di sini: kluster kind-dev namespace default

dev-control-planeBukaPengurus

Tiada dulang di sini.

dev-workerBuka

shop·9448k
shop·q725x

dev-worker2Buka

shop·ddlcm

Masa: +0s (dipercepat)

etcdctl --endpoints=https://127.0.0.1:2379 --cacert=/etc/kubernetes/pki/etcd/ca.crt --cert=/etc/kubernetes/pki/etcd/server.crt --key=/etc/kubernetes/pki/etcd/server.key snapshot save /opt/backup.db
etcdutl --write-out=table snapshot status /opt/backup.db

Teka

Pada etcd 3.7, anda taip etcdctl snapshot status /opt/backup.db. Apa berlaku?

Cuba

Shop dipadam selepas snapshot diambil. pulihkan salinan ke folder baharu yang kosong. Kemudian tunjukkan etcd ke situ. Bahaya: objek yang lebih baharu hilang.

Anda di sini: kluster kind-dev namespace default

dev-control-planeBukaPengurus

Tiada dulang di sini.

dev-workerBuka

Tiada dulang di sini.

dev-worker2Buka

Tiada dulang di sini.

Masa: +0s (dipercepat)

etcdutl --data-dir /var/lib/etcd-restore snapshot restore /opt/backup.db
sed -i 's#/var/lib/etcd$#/var/lib/etcd-restore#' /etc/kubernetes/manifests/etcd.yaml

Semak

Susun langkah pemulihan ikut urutan.

Langkah pertama di atas. Susun ikut turutan. Guna anak panah, atau seret.

Tunggu: etcd mula semula, kemudian API menjawab
kubectl get deployments
etcdutl snapshot status: salinan itu sihat?
Tunjukkan etcd.yaml ke folder baharu
etcdutl --data-dir <folder baharu> snapshot restore

Anda berjaya!

Kini anda boleh sandarkan etcd dan pulihkan objek yang hilang.

Nak tahu lebih (pilihan)

Node kind tiada etcdctl. Di sana, jalankannya dalam pod etcd. Hanya /var/lib/etcd dan sijil dikongsi dengan node itu.

Anda di sini: kluster kind-dev namespace default

~/shop $ kubectl -n kube-system exec etcd-dev-control-plane -- etcdctl --endpoints=https://127.0.0.1:2379 --cacert=/etc/kubernetes/pki/etcd/ca.crt --cert=/etc/kubernetes/pki/etcd/server.crt --key=/etc/kubernetes/pki/etcd/server.key snapshot save /var/lib/etcd/backup.db{"level":"info","ts":"2026-10-31T01:00:22.252272Z","caller":"snapshot/v3_snapshot.go:83","msg":"created temporary db file","path":"/var/lib/etcd/backup.db.part"}{"level":"info","ts":"2026-10-31T01:00:22.255184Z","logger":"client","caller":"v3/maintenance.go:236","msg":"opened snapshot stream; downloading"}{"level":"info","ts":"2026-10-31T01:00:22.258272Z","caller":"snapshot/v3_snapshot.go:96","msg":"fetching snapshot","endpoint":"https://127.0.0.1:2379"}{"level":"info","ts":"2026-10-31T01:00:22.260284Z","logger":"client","caller":"v3/maintenance.go:302","msg":"completed snapshot read; closing"}{"level":"info","ts":"2026-10-31T01:00:22.263372Z","caller":"snapshot/v3_snapshot.go:111","msg":"fetched snapshot","endpoint":"https://127.0.0.1:2379","size":"2.4 MB","took":"9.935898ms","etcd-version":"3.7.0"}{"level":"info","ts":"2026-10-31T01:00:22.266284Z","caller":"snapshot/v3_snapshot.go:121","msg":"saved","path":"/var/lib/etcd/backup.db"}Snapshot saved at /var/lib/etcd/backup.dbServer version 3.7.0

Misi sebenar

Pada kluster kind latihan, simpan snapshot dalam pod etcd (arahan di atas).
kubectl -n kube-system exec etcd-<cluster>-control-plane -- etcdctl --endpoints=https://127.0.0.1:2379 --cacert=/etc/kubernetes/pki/etcd/ca.crt --cert=/etc/kubernetes/pki/etcd/server.crt --key=/etc/kubernetes/pki/etcd/server.key snapshot save /var/lib/etcd/backup.db
Salin keluar dari node, ke laptop anda.
docker cp <cluster>-control-plane:/var/lib/etcd/backup.db ./backup.db

0 daripada 2 langkah siap · Tanda anda disimpan.