github.com/ratanraj/packer@v1.3.2/website/source/docs/extending/custom-builders.html.md (about)

     1  ---
     2  description: |
     3      It is possible to write custom builders using the Packer plugin interface, and
     4      this page documents how to do that.
     5  layout: docs
     6  page_title: 'Custom Builders - Extending'
     7  sidebar_current: 'docs-extending-custom-builders'
     8  ---
     9  
    10  # Custom Builders
    11  
    12  Packer Builders are the components of Packer responsible for creating a machine,
    13  bringing it to a point where it can be provisioned, and then turning that
    14  provisioned machine into some sort of machine image. Several builders are
    15  officially distributed with Packer itself, such as the AMI builder, the VMware
    16  builder, etc. However, it is possible to write custom builders using the Packer
    17  plugin interface, and this page documents how to do that.
    18  
    19  Prior to reading this page, it is assumed you have read the page on [plugin
    20  development basics](/docs/extending/plugins.html).
    21  
    22  ~> **Warning!** This is an advanced topic. If you're new to Packer, we
    23  recommend getting a bit more comfortable before you dive into writing plugins.
    24  
    25  ## The Interface
    26  
    27  The interface that must be implemented for a builder is the `packer.Builder`
    28  interface. It is reproduced below for reference. The actual interface in the
    29  source code contains some basic documentation as well explaining what each
    30  method should do.
    31  
    32  ``` go
    33  type Builder interface {
    34    Prepare(...interface{}) error
    35    Run(ui Ui, hook Hook, cache Cache) (Artifact, error)
    36    Cancel()
    37  }
    38  ```
    39  
    40  ### The "Prepare" Method
    41  
    42  The `Prepare` method for each builder is called prior to any runs with the
    43  configuration that was given in the template. This is passed in as an array of
    44  `interface{}` types, but is generally `map[string]interface{}`. The prepare
    45  method is responsible for translating this configuration into an internal
    46  structure, validating it, and returning any errors.
    47  
    48  For multiple parameters, they should be merged together into the final
    49  configuration, with later parameters overwriting any previous configuration. The
    50  exact semantics of the merge are left to the builder author.
    51  
    52  For decoding the `interface{}` into a meaningful structure, the
    53  [mapstructure](https://github.com/mitchellh/mapstructure) library is
    54  recommended. Mapstructure will take an `interface{}` and decode it into an
    55  arbitrarily complex struct. If there are any errors, it generates very human
    56  friendly errors that can be returned directly from the prepare method.
    57  
    58  While it is not actively enforced, **no side effects** should occur from running
    59  the `Prepare` method. Specifically, don't create files, don't launch virtual
    60  machines, etc. Prepare's purpose is solely to configure the builder and validate
    61  the configuration.
    62  
    63  In addition to normal configuration, Packer will inject a
    64  `map[string]interface{}` with a key of `packer.DebugConfigKey` set to boolean
    65  `true` if debug mode is enabled for the build. If this is set to true, then the
    66  builder should enable a debug mode which assists builder developers and advanced
    67  users to introspect what is going on during a build. During debug builds,
    68  parallelism is strictly disabled, so it is safe to request input from stdin and
    69  so on.
    70  
    71  ### The "Run" Method
    72  
    73  `Run` is where all the interesting stuff happens. Run is executed, often in
    74  parallel for multiple builders, to actually build the machine, provision it, and
    75  create the resulting machine image, which is returned as an implementation of
    76  the `packer.Artifact` interface.
    77  
    78  The `Run` method takes three parameters. These are all very useful. The
    79  `packer.Ui` object is used to send output to the console. `packer.Hook` is used
    80  to execute hooks, which are covered in more detail in the hook section below.
    81  And `packer.Cache` is used to store files between multiple Packer runs, and is
    82  covered in more detail in the cache section below.
    83  
    84  Because builder runs are typically a complex set of many steps, the
    85  [multistep](https://github.com/hashicorp/packer/blob/master/helper/multistep)
    86  helper is recommended to bring order to the complexity. Multistep is a library
    87  which allows you to separate your logic into multiple distinct "steps" and
    88  string them together. It fully supports cancellation mid-step and so on. Please
    89  check it out, it is how the built-in builders are all implemented.
    90  
    91  Finally, as a result of `Run`, an implementation of `packer.Artifact` should be
    92  returned. More details on creating a `packer.Artifact` are covered in the
    93  artifact section below. If something goes wrong during the build, an error can
    94  be returned, as well. Note that it is perfectly fine to produce no artifact and
    95  no error, although this is rare.
    96  
    97  ### The "Cancel" Method
    98  
    99  The `Run` method is often run in parallel. The `Cancel` method can be called at
   100  any time and requests cancellation of any builder run in progress. This method
   101  should block until the run actually stops.
   102  
   103  Cancels are most commonly triggered by external interrupts, such as the user
   104  pressing `Ctrl-C`. Packer will only exit once all the builders clean up, so it
   105  is important that you architect your builder in a way that it is quick to
   106  respond to these cancellations and clean up after itself.
   107  
   108  ## Creating an Artifact
   109  
   110  The `Run` method is expected to return an implementation of the
   111  `packer.Artifact` interface. Each builder must create their own implementation.
   112  The interface has ample documentation to help you get started.
   113  
   114  The only part of an artifact that may be confusing is the `BuilderId` method.
   115  This method must return an absolutely unique ID for the builder. In general, I
   116  follow the practice of making the ID contain my GitHub username and then the
   117  platform it is building for. For example, the builder ID of the VMware builder
   118  is "hashicorp.vmware" or something similar.
   119  
   120  Post-processors use the builder ID value in order to make some assumptions about
   121  the artifact results, so it is important it never changes.
   122  
   123  Other than the builder ID, the rest should be self-explanatory by reading the
   124  [packer.Artifact interface
   125  documentation](https://github.com/hashicorp/packer/blob/master/packer/artifact.go).
   126  
   127  ## Provisioning
   128  
   129  Packer has built-in support for provisioning, but the moment when provisioning
   130  runs must be invoked by the builder itself, since only the builder knows when
   131  the machine is running and ready for communication.
   132  
   133  When the machine is ready to be provisioned, run the `packer.HookProvision`
   134  hook, making sure the communicator is not nil, since this is required for
   135  provisioners. An example of calling the hook is shown below:
   136  
   137  ``` go
   138  hook.Run(packer.HookProvision, ui, comm, nil)
   139  ```
   140  
   141  At this point, Packer will run the provisioners and no additional work is
   142  necessary.
   143  
   144  -> **Note:** Hooks are still undergoing thought around their general design
   145  and will likely change in a future version. They aren't fully "baked" yet, so
   146  they aren't documented here other than to tell you how to hook in provisioners.
   147  
   148  ## Caching Files
   149  
   150  It is common for some builders to deal with very large files, or files that take
   151  a long time to generate. For example, the VMware builder has the capability to
   152  download the operating system ISO from the internet. This is timely process, so
   153  it would be convenient to cache the file. This sort of caching is a core part of
   154  Packer that is exposed to builders.
   155  
   156  The cache interface is `packer.Cache`. It behaves much like a Go
   157  [RWMutex](https://golang.org/pkg/sync/#RWMutex). The builder requests a "lock" on
   158  certain cache keys, and is given exclusive access to that key for the duration
   159  of the lock. This locking mechanism allows multiple builders to share cache data
   160  even though they're running in parallel.
   161  
   162  For example, both the VMware and VirtualBox builders support downloading an
   163  operating system ISO from the internet. Most of the time, this ISO is identical.
   164  The locking mechanisms of the cache allow one of the builders to download it
   165  only once, but allow both builders to share the downloaded file.
   166  
   167  The [documentation for
   168  packer.Cache](https://github.com/hashicorp/packer/blob/master/packer/cache.go)
   169  is very detailed in how it works.