Manually create a backup of a healthy etcd cluster and check to see if the backup was created successfully.
Backups of healthy etcd clusters can be used to restore the cluster if it becomes unhealthy at any point.
The commands in this procedure can be run on any master node (ncn-mXXX) or worker node (ncn-wXXX) on the system.
A healthy etcd cluster is available on the system. See Check the Health of etcd Clusters.
Create a backup for the desired etcd cluster.
The example below is backing up the etcd cluster for the Boot Script Service (BSS) named wednesday-manual-backup.
/opt/cray/platform-utils/etcd/etcd-util.sh create_backup cray-bss wednesday-manual-backup
Example output:
Taking snapshot from cray-bss-bitnami-etcd-0...
Pushing newly created snapshot /snapshots/cray-bss-bitnami-etcd/db-2023-03-10_23-38 to S3 as wednesday-manual-backup for cray-bss
upload: snapshots/cray-bss-bitnami-etcd/db-2023-03-10_23-38 to s3://etcd-backup/cray-bss/wednesday-manual-backup
Verify the newly created backup is available in S3:
/opt/cray/platform-utils/etcd/etcd-util.sh list_backups cray-bss
Example output:
cray-bss/db-2023-03-10_21-00
cray-bss/db-2023-03-10_22-00
cray-bss/db-2023-03-10_23-00
cray-bss/wednesday-manual-backup