User data

Adding a user and SSH key

The following cloud-init example shows how a Linux user can add a user and provide the user with an authorized SSH key. The Name field has the public key that is added to ~/.ssh/authorized_keys.

For Linux images, the Ed25519 SSH key type can be used only if the SSH server for the operating system supports this key type. For more information, see Getting started with SSH keys.

#cloud-config
users:
  - name: demouser
    gecos: Demo User
    sudo: ALL=(ALL) NOPASSWD:ALL
    groups: users, admin
    ssh_import_id: None
    lock_passwd: true
    ssh_authorized_keys:
        - <ssh public key>

The following shell script example shows how a Linux user can add an SSH key for the current user.

#!/bin/bash
echo <sshKey> > ~/.ssh/authorized_keys

You can paste one of these examples directly into the User Data field. The user data is then available to the virtual server instance during provisioning.

If you specify to include a file and have spaces preceding the file name, the data isn't interpreted correctly. Verify that #!/bin/sh or #!/bin/bash are the first characters on the line immediately following the end of file designation (<<EOF). The characters can't be indented.

For more Linux user data examples and information, see Cloud config examples).

Configuring a single disk instance storage by using cloud-config script

Instance storage is a feature of VPC where you can request virtual server instances with fast, local storage attached. For more information about instance storage, refer to About instance storage.

By default, when you provision a virtual server instance with instance storage and then log in to that server for the first time, the instance storage disks are not configured. They show up as block devices (for example /dev/vdb), and need the following done before you can use them with a file system:

• Partition the device.
• Format the partition with a file system.
• Mount the file system.

These activities can also be performed automatically when you use the User Data field to provision the virtual instance. The cloud-config script defines these actions:

• Instructs the cloud initialization process to run during first boot of your virtual instance.
• Automatically partitions the device.
• Formats the partition with an ext4 file system.
• Mounts the file system.

Do not specify the User Data field to automatically configure an instance storage disk that is not part of the virtual instance definition. Attempting to automatically configure an instance storage disk that is not part of the virtual instance definition would target dev/vdb device and might point to the cloud-init data source, which would corrupt the cloud-init configurations such as hostname and ssh keys.

The following example shows user data that automatically configures an instance storage disk. This example can be used with an instance profile that specifies a single disk.

#cloud-config
# Cloud-init supports simple partition and file system config.
# This user data yaml will create a full partition on the first
# virtio_blk device after the boot device, initialize ext4 on it and
# mount it on a folder matching the label.
#
disk_setup:
  /dev/vdb:
    table_type: 'mbr'
    layout:
    - 100
    overwrite: false

fs_setup:
  - label: /mnt/inststg1
    filesystem: 'ext4'
    device: /dev/vdb1
    overwrite: false

runcmd:
  - [ mkdir, /mnt/inststg1 ]

mounts:
  - ["/dev/vdb1", "/mnt/inststg1"]

mount_default_fields: [ None, None, "auto", "defaults,nofail", "0", "2" ]

This script configures /dev/vdb, the first instance storage device with virtio_blk interface type available on the virtual instance. This script can be pasted into the User Data field or imported by using the Import user data link on the UI.

This script was tested with a stock image of Ubuntu 20.04 LTS Focal Fossa. While the script might be appropriate for other Linux stock and custom images, it is possible that adjustments to the script might be required.

The cloud-config script is not appropriate for Windows virtual servers. If your instance does not contain instance storage, do not use this script. It might configure an unintended device.

Here are some details about items that are found in the cloud-config script:

• disk_setup: creates a single default partition spanning the entire disk with a master boot record. The “overwrite: false” causes this operation to be skipped if the device is already partitioned.
• fs_setup: creates an ext4 file system on the first partition of the vdb block device and gives it the label “inststg1”. You can adjust the file system type and label to suit your needs. The “overwrite: false” setting prevents formatting the partition if it is already formatted with a file system.
• runcmd: creates a directory to use as a mount-point off ‘/’ in the hierarchical parent file system.
• mounts: performs the mount of the file system device onto /inststg1. You can rename the directory path here and under runcmd.
• mount_default_fields: creates a permanent mount directive in the /etc/fstab file. If you issue a reboot command, the virtual server comes back up with the instance storage file system still mounted.

The CLI and API also supports the User Data field.

This cloud-config script example automatically configures an instance storage /dev/vdb block device. If you only reboot your virtual server, this configuration continues to work since the configuration and data persists. However, if you start a virtual server that was previously stopped, then the virtual server has a fresh set of instance storage disks when it boots up. This situation would require the cloud-init steps to be manually run again. By default, the cloud-config script is only run on first boot. You can also edit the cloud-init section of the cloud.config file so that the cloud-init steps are automatically run on each boot. For the steps, see the Edit the cloud_cloud_init_modules section of the cloud.config file to run on each boot section.

Configuring a two disk instance storage by using cloud-config script

The following example shows user data that automatically configures an instance storage disk. This example can be used with an instance profile that specifies two disks.

#cloud-config
# Cloud-init supports simple partition and file system config.
# This user data yaml will create a full partition on the first two
# virtio_blk devices after the boot device, initialize ext4 on them and
# mount them on new folders off of ‘/mnt/’.
#
disk_setup:
  /dev/vdb:
    table_type: 'mbr'
    layout:
    - 100
    overwrite: false
  /dev/vdc:
    table_type: 'mbr'
    layout:
    - 100
    overwrite: false

fs_setup:
  - label: /mnt/inststg1
    filesystem: 'ext4'
    device: /dev/vdb1
    overwrite: false
  - label: /mnt/inststg2
    filesystem: 'ext4'
    device: /dev/vdc1
    overwrite: false

runcmd:
  - [ mkdir, /mnt/inststg1 ]
  - [ mkdir, /mnt/inststg2 ]

mounts:
  - ["/dev/vdb1", "/mnt/inststg1"]
  - ["/dev/vdc1", "/mnt/inststg2"]

mount_default_fields: [ None, None, "auto", "defaults,nofail", "0", "2" ]

This cloud-config script example automatically configures both the instance storage /dev/vdb and dev/vdc block devices, and assumes a virtio_blk interface type for instance storage. If you only reboot your virtual server, this configuration continues to work since the configuration and data persists.

If you start a virtual server that was previously stopped, then the virtual server has a fresh set of instance storage disks when it boots up. This situation requires the cloud-init steps to be manually run again. By default, the cloud-config script is only run on first boot. You can also edit the cloud-init section of the cloud.config file so that the cloud-init steps are automatically run on each boot. For more information, see Edit the cloud_cloud_init_modules section of the cloud.config file to run on each boot.