packer-cn/website/source/docs/builders/amazon-instance.html.md

22 KiB

description layout page_title sidebar_current
The amazon-instance Packer builder is able to create Amazon AMIs backed by instance storage as the root device. For more information on the difference between instance storage and EBS-backed instances, see the storage for the root device section in the EC2 documentation. docs Amazon instance-store - Builders docs-builders-amazon-instance

AMI Builder (instance-store)

Type: amazon-instance

The amazon-instance Packer builder is able to create Amazon AMIs backed by instance storage as the root device. For more information on the difference between instance storage and EBS-backed instances, see the "storage for the root device" section in the EC2 documentation.

This builder builds an AMI by launching an EC2 instance from an existing instance-storage backed AMI, provisioning that running machine, and then bundling and creating a new AMI from that machine. This is all done in your own AWS account. The builder will create temporary key pairs, security group rules, etc. that provide it temporary access to the instance while the image is being created. This simplifies configuration quite a bit.

The builder does not manage AMIs. Once it creates an AMI and stores it in your account, it is up to you to use, delete, etc. the AMI.

-> Note: Temporary resources are, by default, all created with the prefix packer. This can be useful if you want to restrict the security groups and key pairs packer is able to operate on.

-> Note: This builder requires that the Amazon EC2 AMI Tools are installed onto the machine. This can be done within a provisioner, but must be done before the builder finishes running.

~> Instance builds are not supported for Windows. Use amazon-ebs instead.

Configuration Reference

There are many configuration options available for the builder. They are segmented below into two categories: required and optional parameters. Within each category, the available configuration keys are alphabetized.

In addition to the options listed here, a communicator can be configured for this builder.

Required:

  • access_key (string) - The access key used to communicate with AWS. Learn how to set this.

  • account_id (string) - Your AWS account ID. This is required for bundling the AMI. This is not the same as the access key. You can find your account ID in the security credentials page of your AWS account.

  • ami_name (string) - The name of the resulting AMI that will appear when managing AMIs in the AWS console or via APIs. This must be unique. To help make this unique, use a function like timestamp (see configuration templates for more info)

  • instance_type (string) - The EC2 instance type to use while building the AMI, such as m1.small.

  • region (string) - The name of the region, such as us-east-1, in which to launch the EC2 instance to create the AMI.

  • s3_bucket (string) - The name of the S3 bucket to upload the AMI. This bucket will be created if it doesn't exist.

  • secret_key (string) - The secret key used to communicate with AWS. Learn how to set this.

  • source_ami (string) - The initial AMI used as a base for the newly created machine.

  • x509_cert_path (string) - The local path to a valid X509 certificate for your AWS account. This is used for bundling the AMI. This X509 certificate must be registered with your account from the security credentials page in the AWS console.

  • x509_key_path (string) - The local path to the private key for the X509 certificate specified by x509_cert_path. This is used for bundling the AMI.

Optional:

  • ami_block_device_mappings (array of block device mappings) - Add one or more block device mappings to the AMI. These will be attached when booting a new instance from your AMI. To add a block device during the Packer build see launch_block_device_mappings below. Your options here may vary depending on the type of VM you use. The block device mappings allow for the following configuration:

    • delete_on_termination (boolean) - Indicates whether the EBS volume is deleted on instance termination. Default false. NOTE: If this value is not explicitly set to true and volumes are not cleaned up by an alternative method, additional volumes will accumulate after every build.

    • device_name (string) - The device name exposed to the instance (for example, /dev/sdh or xvdh). Required when specifying volume_size.

    • encrypted (boolean) - Indicates whether to encrypt the volume or not

    • iops (number) - The number of I/O operations per second (IOPS) that the volume supports. See the documentation on IOPs for more information

    • no_device (boolean) - Suppresses the specified device included in the block device mapping of the AMI

    • snapshot_id (string) - The ID of the snapshot

    • virtual_name (string) - The virtual device name. See the documentation on Block Device Mapping for more information

    • volume_size (number) - The size of the volume, in GiB. Required if not specifying a snapshot_id

    • volume_type (string) - The volume type. gp2 for General Purpose (SSD) volumes, io1 for Provisioned IOPS (SSD) volumes, and standard for Magnetic volumes

  • ami_description (string) - The description to set for the resulting AMI(s). By default this description is empty. This is a template engine where the SourceAMI variable is replaced with the source AMI ID and BuildRegion variable is replaced with the value of region.

  • ami_groups (array of strings) - A list of groups that have access to launch the resulting AMI(s). By default no groups have permission to launch the AMI. all will make the AMI publicly accessible. AWS currently doesn't accept any value other than all.

  • ami_product_codes (array of strings) - A list of product codes to associate with the AMI. By default no product codes are associated with the AMI.

  • ami_regions (array of strings) - A list of regions to copy the AMI to. Tags and attributes are copied along with the AMI. AMI copying takes time depending on the size of the AMI, but will generally take many minutes.

  • ami_users (array of strings) - A list of account IDs that have access to launch the resulting AMI(s). By default no additional users other than the user creating the AMI has permissions to launch it.

  • ami_virtualization_type (string) - The type of virtualization for the AMI you are building. This option is required to register HVM images. Can be paravirtual (default) or hvm.

  • associate_public_ip_address (boolean) - If using a non-default VPC, public IP addresses are not provided by default. If this is toggled, your new instance will get a Public IP.

  • availability_zone (string) - Destination availability zone to launch instance in. Leave this empty to allow Amazon to auto-assign.

  • bundle_destination (string) - The directory on the running instance where the bundled AMI will be saved prior to uploading. By default this is /tmp. This directory must exist and be writable.

  • bundle_prefix (string) - The prefix for files created from bundling the root volume. By default this is image-{{timestamp}}. The timestamp variable should be used to make sure this is unique, otherwise it can collide with other created AMIs by Packer in your account.

  • bundle_upload_command (string) - The command to use to upload the bundled volume. See the "custom bundle commands" section below for more information.

  • bundle_vol_command (string) - The command to use to bundle the volume. See the "custom bundle commands" section below for more information.

  • custom_endpoint_ec2 (string) - This option is useful if you use a cloud provider whose API is compatible with aws EC2. Specify another endpoint like this https://ec2.custom.endpoint.com.

  • ebs_optimized (boolean) - Mark instance as EBS Optimized. Default false.

  • ena_support (boolean) - Enable enhanced networking (ENA but not SriovNetSupport) on HVM-compatible AMIs. If true, add ec2:ModifyInstanceAttribute to your AWS IAM policy. Note: you must make sure enhanced networking is enabled on your instance. See Amazon's documentation on enabling enhanced networking. Default false.

  • force_deregister (boolean) - Force Packer to first deregister an existing AMI if one with the same name already exists. Defaults to false.

  • force_delete_snapshot (boolean) - Force Packer to delete snapshots associated with AMIs, which have been deregistered by force_deregister. Defaults to false.

  • iam_instance_profile (string) - The name of an IAM instance profile to launch the EC2 instance with.

  • launch_block_device_mappings (array of block device mappings) - Add one or more block devices before the Packer build starts. These are not necessarily preserved when booting from the AMI built with Packer. See ami_block_device_mappings, above, for details.

  • mfa_code (string) - The MFA TOTP code. This should probably be a user variable since it changes all the time.

  • profile (string) - The profile to use in the shared credentials file for AWS. See Amazon's documentation on specifying profiles for more details.

  • region_kms_key_ids (map of strings) - a map of regions to copy the ami to, along with the custom kms key id to use for encryption for that region. Keys must match the regions provided in ami_regions. If you just want to encrypt using a default ID, you can stick with kms_key_id and ami_regions. If you want a region to be encrypted with that region's default key ID, you can use an empty string "" instead of a key id in this map. (e.g. "us-east-1": "") However, you cannot use default key IDs if you are using this in conjunction with snapshot_users -- in that situation you must use custom keys.

  • run_tags (object of key/value strings) - Tags to apply to the instance that is launched to create the AMI. These tags are not applied to the resulting AMI unless they're duplicated in tags. This is a template engine where the SourceAMI variable is replaced with the source AMI ID and BuildRegion variable is replaced with the value of region.

  • security_group_id (string) - The ID (not the name) of the security group to assign to the instance. By default this is not set and Packer will automatically create a new temporary security group to allow SSH access. Note that if this is specified, you must be sure the security group allows access to the ssh_port given below.

  • security_group_ids (array of strings) - A list of security groups as described above. Note that if this is specified, you must omit the security_group_id.

  • temporary_security_group_source_cidr (string) - An IPv4 CIDR block to be authorized access to the instance, when packer is creating a temporary security group. The default is 0.0.0.0/0 (ie, allow any IPv4 source). This is only used when security_group_id or security_group_ids is not specified.

  • skip_region_validation (boolean) - Set to true if you want to skip validation of the region configuration option. Defaults to false.

  • snapshot_groups (array of strings) - A list of groups that have access to create volumes from the snapshot(s). By default no groups have permission to create volumes form the snapshot(s). all will make the snapshot publicly accessible.

  • snapshot_users (array of strings) - A list of account IDs that have access to create volumes from the snapshot(s). By default no additional users other than the user creating the AMI has permissions to create volumes from the backing snapshot(s).

  • source_ami_filter (object) - Filters used to populate the source_ami field. Example:

    {
      "source_ami_filter": {
        "filters": {
          "virtualization-type": "hvm",
          "name": "ubuntu/images/*ubuntu-xenial-16.04-amd64-server-*",
          "root-device-type": "ebs"
        },
        "owners": ["099720109477"],
        "most_recent": true
      }
    }
    

    This selects the most recent Ubuntu 16.04 HVM EBS AMI from Canonical. NOTE: This will fail unless exactly one AMI is returned. In the above example, most_recent will cause this to succeed by selecting the newest image.

    • filters (map of strings) - filters used to select a source_ami. NOTE: This will fail unless exactly one AMI is returned. Any filter described in the docs for DescribeImages is valid.

    • owners (array of strings) - This scopes the AMIs to certain Amazon account IDs. This is helpful to limit the AMIs to a trusted third party, or to your own account.

    • most_recent (boolean) - Selects the newest created image when true. This is most useful for selecting a daily distro build.

  • snapshot_tags (object of key/value strings) - Tags to apply to snapshot. They will override AMI tags if already applied to snapshot.

  • spot_price (string) - The maximum hourly price to launch a spot instance to create the AMI. It is a type of instances that EC2 starts when the maximum price that you specify exceeds the current spot price. Spot price will be updated based on available spot instance capacity and current spot Instance requests. It may save you some costs. You can set this to auto for Packer to automatically discover the best spot price or to 0 to use an on-demand instance (default).

  • spot_price_auto_product (string) - Required if spot_price is set to auto. This tells Packer what sort of AMI you're launching to find the best spot price. This must be one of: Linux/UNIX, SUSE Linux, Windows, Linux/UNIX (Amazon VPC), SUSE Linux (Amazon VPC), Windows (Amazon VPC)

  • sriov_support (boolean) - Enable enhanced networking (SriovNetSupport but not ENA) on HVM-compatible AMIs. If true, add ec2:ModifyInstanceAttribute to your AWS IAM policy. Note: you must make sure enhanced networking is enabled on your instance. See Amazon's documentation on enabling enhanced networking. Default false.

  • ssh_keypair_name (string) - If specified, this is the key that will be used for SSH with the machine. The key must match a key pair name loaded up into Amazon EC2. By default, this is blank, and Packer will generate a temporary key pair unless ssh_password is used. ssh_private_key_file or ssh_agent_auth must be specified when ssh_keypair_name is utilized.

  • ssh_agent_auth (boolean) - If true, the local SSH agent will be used to authenticate connections to the source instance. No temporary key pair will be created, and the values of ssh_password and ssh_private_key_file will be ignored. To use this option with a key pair already configured in the source AMI, leave the ssh_keypair_name blank. To associate an existing key pair in AWS with the source instance, set the ssh_keypair_name field to the name of the key pair.

  • ssh_private_ip (boolean) - If true, then SSH will always use the private IP if available. Also works for WinRM.

  • subnet_id (string) - If using VPC, the ID of the subnet, such as subnet-12345def, where Packer will launch the EC2 instance. This field is required if you are using an non-default VPC.

  • tags (object of key/value strings) - Tags applied to the AMI. This is a template engine where the SourceAMI variable is replaced with the source AMI ID and BuildRegion variable is replaced with the value of region.

  • temporary_key_pair_name (string) - The name of the temporary key pair to generate. By default, Packer generates a name that looks like packer_<UUID>, where <UUID> is a 36 character unique identifier.

  • user_data (string) - User data to apply when launching the instance. Note that you need to be careful about escaping characters due to the templates being JSON. It is often more convenient to use user_data_file, instead.

  • user_data_file (string) - Path to a file that will be used for the user data when launching the instance.

  • vpc_id (string) - If launching into a VPC subnet, Packer needs the VPC ID in order to create a temporary security group within the VPC. Requires subnet_id to be set. If this field is left blank, Packer will try to get the VPC ID from the subnet_id.

  • x509_upload_path (string) - The path on the remote machine where the X509 certificate will be uploaded. This path must already exist and be writable. X509 certificates are uploaded after provisioning is run, so it is perfectly okay to create this directory as part of the provisioning process. Defaults to /tmp.

  • windows_password_timeout (string) - The timeout for waiting for a Windows password for Windows instances. Defaults to 20 minutes. Example value: 10m

Basic Example

Here is a basic example. It is completely valid except for the access keys:

{
  "type": "amazon-instance",
  "access_key": "YOUR KEY HERE",
  "secret_key": "YOUR SECRET KEY HERE",
  "region": "us-east-1",
  "source_ami": "ami-d9d6a6b0",
  "instance_type": "m1.small",
  "ssh_username": "ubuntu",

  "account_id": "0123-4567-0890",
  "s3_bucket": "packer-images",
  "x509_cert_path": "x509.cert",
  "x509_key_path": "x509.key",
  "x509_upload_path": "/tmp",

  "ami_name": "packer-quick-start {{timestamp}}"
}

-> Note: Packer can also read the access key and secret access key from environmental variables. See the configuration reference in the section above for more information on what environmental variables Packer will look for.

Accessing the Instance to Debug

If you need to access the instance to debug for some reason, run the builder with the -debug flag. In debug mode, the Amazon builder will save the private key in the current directory and will output the DNS or IP information as well. You can use this information to access the instance as it is running.

Custom Bundle Commands

A lot of the process required for creating an instance-store backed AMI involves commands being run on the actual source instance. Specifically, the ec2-bundle-vol and ec2-upload-bundle commands must be used to bundle the root filesystem and upload it, respectively.

Each of these commands have a lot of available flags. Instead of exposing each possible flag as a template configuration option, the instance-store AMI builder for Packer lets you customize the entire command used to bundle and upload the AMI.

These are configured with bundle_vol_command and bundle_upload_command. Both of these configurations are configuration templates and have support for their own set of template variables.

Bundle Volume Command

The default value for bundle_vol_command is shown below. It is split across multiple lines for convenience of reading. The bundle volume command is responsible for executing ec2-bundle-vol in order to store and image of the root filesystem to use to create the AMI.

sudo -i -n ec2-bundle-vol \
  -k {{.KeyPath}}  \
  -u {{.AccountId}} \
  -c {{.CertPath}} \
  -r {{.Architecture}} \
  -e {{.PrivatePath}}/* \
  -d {{.Destination}} \
  -p {{.Prefix}} \
  --batch \
  --no-filter

The available template variables should be self-explanatory based on the parameters they're used to satisfy the ec2-bundle-vol command.

~> Warning! Some versions of ec2-bundle-vol silently ignore all .pem and .gpg files during the bundling of the AMI, which can cause problems on some systems, such as Ubuntu. You may want to customize the bundle volume command to include those files (see the --no-filter option of ec2-bundle-vol).

Bundle Upload Command

The default value for bundle_upload_command is shown below. It is split across multiple lines for convenience of reading. Access key and secret key are omitted if using instance profile. The bundle upload command is responsible for taking the bundled volume and uploading it to S3.

sudo -i -n ec2-upload-bundle \
  -b {{.BucketName}} \
  -m {{.ManifestPath}} \
  -a {{.AccessKey}} \
  -s {{.SecretKey}} \
  -d {{.BundleDirectory}} \
  --batch \
  --region {{.Region}} \
  --retry

The available template variables should be self-explanatory based on the parameters they're used to satisfy the ec2-upload-bundle command. Additionally, {{.Token}} is available when overriding this command. You must create your own bundle command with the addition of -t {{.Token}} if you are assuming a role.