Wipe Drives

WARNING: This is the point of no return. Once the disks are wiped, the node must be rebuilt.

All commands in this section must be run on the node being rebuilt (unless otherwise indicated). These commands can be done from the ConMan console window.

Only follow the steps in the section for the node type that is being rebuilt:

Wipe Disks: Master

  1. Unmount the etcd volume and remove the volume group.

    NOTE: etcd should already be stopped as part of the “Prepare Master Node” steps.

    /run/lib-etcd
vgremove -f etcdvg0-ETCDK8S

  2. Unmount the SDU mountpoint and remove the volume group.

    umount /var/lib/sdu
vgremove -f metalvg0-CRAYSDU

  3. Wipe the drives

    mdisks=$(lsblk -l -o SIZE,NAME,TYPE,TRAN | grep -E '(sata|nvme|sas)' | sort -h | awk '   {print "/dev/" $2}')
wipefs --all --force $mdisks

Wipe Disks: Worker Node

  1. Stop containerd and wipe drives.

    systemctl stop containerd.service

  2. Unmount partitions.

    umount /var/lib/kubelet
umount /run/lib-containerd
umount /run/containerd

  3. Wipe the drives.

    wipefs --all --force /dev/sd* /dev/disk/by-label/*

Wipe Disks: Utility Storage Node

  1. Stop running OSDs on the node being wiped.

    ncn-s# systemctl stop ceph-osd.target

  2. Make sure the OSDs (if any) are not running after running the first command.

    ncn-s# ls -1 /dev/sd* /dev/disk/by-label/*
ncn-s# vgremove -f --select 'vg_name=~ceph*'

  3. Unmount and remove the metalvg0 volume group.

    umount /etc/ceph
umount /var/lib/ceph
umount /var/lib/containers
vgremove -f metalvg0

  4. Wipe the disks and RAIDs.

    wipefs --all --force /dev/sd* /dev/disk/by-label/*

Next Step

Proceed to the next step to Power Cycle and Rebuild Nodes or return to the main Rebuild NCNs page.