Welcome to ploy’s documentation!¶

ploy¶

Overview¶

Ploy is a commandline-tool to provision, manage and control server instances. There are plugins for EC2 (ploy_ec2), FreeBSD Jails (ploy_ezjail) and more. You can create, delete, monitor and ssh into instances while ploy handles the details like ssh fingerprint checking. Additional plugins provide advanced functionality like integrating Fabric (ploy_fabric) and Ansible (ploy_ansible).

You can find the detailed documentation at http://ploy.readthedocs.org/en/latest/

Installation¶

ploy is best installed with easy_install, pip or with zc.recipe.egg in a buildout. It installs two scripts, ploy and ploy-ssh.

Configuration¶

All information about server instances is located in ploy.conf, which by default is looked up in etc/ploy.conf.

Plugins¶

Support for backends and further functionality is implemented by plugins. One plugin is included with ploy.

plain: For regular servers accessible via ssh.

You can see which plugins are available in your current installation with ploy -v.

Plain¶

With plain instances you can put infos about servers into the configuration to benefit from some ploy features like ssh fingerprint checking and plugins like the Fabric integration.

Options¶

host or ip

(required) The host name or address for the server.

user

The default user for ssh connections. If it’s set to * then the current local user name is used.

port

The ssh port number.

ssh-host-keys or ssh-fingerprints

(required) The public ssh host keys (or fingerprints) of the server.

ssh-host-keys

It’s best to use the public ssh host keys if available. You can either specify the path, or put in the contents directly (ssh-rsa AAAA...). Multiple host keys can be specified one per line, or separated by commas.

ssh-fingerprints

If set to ask then manual interactive verification is enabled. If set to ignore then no verification is performed at all! Multiple fingerprints can be specified one per line, or separated by commas. The format of fingerprints is either 16 bytes as hex numbers separated by colons, the same 16 bytes prefixed by MD5:, or the hash type, followed by a colon and the base 64 encoded hash digest.

password-fallback

If this boolean is true, then using a password as fallback is enabled if the ssh key doesn’t work. This is off by default. You may be asked more than once for the password. The first time is by paramiko which always happens, but is remembered. The other times is by the ssh command line tool if it’s invoked.

password

Never use this directly! If password-fallback is enabled this password is used. This is mainly meant for Fabric scripts which have other ways to get the password. On use case is bootstrapping FreeBSD from an mfsBSD distribution where the password is fixed.

proxycommand

The command to use in the ProxyCommand option for ssh when using the ploy-ssh command. There are some variables which can be used:

path

The directory of the ploy.conf file. Useful if you want to use the ploy-ssh command itself for the proxy.

known_hosts

The absolute path to the known_hosts file managed by ploy.

instances

The variables of other instances. For example: instances.foo.ip

In addition to these the variables of the instance itself are available.

A full example for a proxycommand:

proxycommand = {path}/../bin/ploy-ssh vm-master -W {ip}:22

ssh-key-filename

Location of private ssh key to use.

ssh-extra-args

A list of settings separated by newlines passed on to ssh.

Example:

ssh-extra-args = ForwardAgent yes

SSH integration¶

ploy provides an additional tool ploy-ssh to easily perform SSH based operations against named instances. Particularly, it encapsulates the entire SSH fingerprint mechanism. For example EC2 instances are often short-lived and normally trigger warnings, especially, if you are using elastic IPs.

Note:: it does so not by simply turning off these checks, but by transparently updating its own fingerprint list using mechanisms provided by the backend plugins.

The easiest scenario is simply to create an SSH session with an instance. You can either use the ssh subcommand of the ploy tool like so:

ploy ssh INSTANCENAME

Alternatively you can use the ploy-ssh command directly, like so:

ploy-ssh INSTANCENAME

The latter has been provided to support scp and rsync. Here are some examples, you get the idea:

scp -S `pwd`/bin/ploy-ssh some.file demo-server:/some/path/
rsync -e "bin/ploy-ssh" some/path fschulze@demo-server:/some/path

Instance names¶

Instances have an id which is the part after the colon in the configuration. They also have a unique id which has the form [masterid]-[instanceid]. The [masterid] depends on the plugin. For plain instances it is plain. The [instanceid] is the id of the instance. So, if you have the following config:

[plain-instance:foo]
...

Then the unique id of the instance is plain-foo.

Macro expansion¶

In the ploy.conf you can use macro expansion for cleaner configuration files. That looks like this:

[ec2-instance:demo-server2]
<= demo-server
securitygroups = demo-server2

[ec2-securitygroup:demo-server2]
<= demo-server

All the options from the specified macro are copied with some exceptions depending on the backend plugin.

If you want to copy data from some other kind of options, you can add a colon in the macro name. This is useful if you want to have a base for instances like this:

[macro:base-instance]
keypair = default
region = eu-west-1
placement = eu-west-1a

[ec2-instance:server]
<= macro:base-instance
...

Massaging of config values¶

Plugins and ploy massage certain string values from the config to convert them to other types and do formatting or expansion.

You can use that yourself, which is useful for the Fabric integration and other things.

Here is a simple example:

[section]
massagers =
  intvalue=ploy.config.IntegerMassager
  boolvalue=ploy.config.BooleanMassager
intvalue = 1
boolvalue = yes

If you now access those values from for example a fabric task, you get the correct type instead of strings.

The above syntax registers the massagers only for that section. You can register massagers for other sections or even section groups with this syntax:

massagers =
  [option]=[sectiongroup]:import.path.to.massager
  [option]=[sectiongroup]:[section]:import.path.to.massager

The parts have the following meaning:

[option]

This is the name of the option which should be massaged

[sectiongroup]

The name of the section group. That’s the part before the optional colon in a section. To match sections without a colon, use global. To match every section, use *.

[section]

The name of the section to which this massager is applied. If empty, the current section is used.

Buildout specifics¶

With zc.recipe.egg you can set a custom configfile location like this:

[ploy]
recipe = zc.recipe.egg
eggs = ploy
arguments = configpath="${buildout:directory}/etc/", configname="servers.cfg"

Changelog¶

1.5.1 - 2017-12-17¶

Fix getting fingerprints automatically when ssh-host-keys is used. [fschulze]

1.5.0 - 2017-10-03¶

New ssh-host-keys option which allows to set host keys directly to force paramiko to use a specific key type to avoid fingerprint mismatches. [fschulze]

1.4.0 - 2017-10-02¶

Allow instance implementations to return multiple fingerprints via get_fingerprints method. [fschulze]
Support other key types than just rsa, because paramiko > 2 defaults to ed25519. [fschulze]
Output more information about used keys on fingerprint mismatch. [fschulze]

1.3.1 - 2016-06-02¶

Don’t get ssh-fingerprints from master if not set in instance config. This fixes automatic fingerprint fetching for EC2, ezjail etc. [fschulze]

1.3.0 - 2016-05-21¶

Add option ssh-fingerprints which allows to specify multiple fingerprints. [fschulze]
Support new output of ssh-keygen which includes the hash type and defaults to SHA256. [fschulze]

1.2.1 - 2015-08-27¶

Allow to specify multiple masters per instance. [fschulze]

1.2.0 - 2015-03-05¶

Add Executor helper to handle local and remote command execution. It’s also handling ssh agent forwarding enabled by either the users ssh config or the ssh-extra-args option. [fschulze]

1.1.0 - 2015-02-28¶

Add ssh-extra-args option. [fschulze]
Add annotate command to print the configuration with the source of each setting. [fschulze]
Allow custom shebang in gzipped startup scripts. [fschulze]

1.0.3 - 2015-01-22¶

Drop bad entries from our known_hosts file to prevent failures in paramiko. [fschulze]
Set StrictHostKeyChecking=yes for all ssh connections to prevent interactive asking. [fschulze]

1.0.2 - 2014-10-04¶

Ask before terminating an instance. [fschulze]
Fix config setting propagation in some cases of proxied instances. [fschulze]
Close all connections before exiting. This prevents hangs caused by open proxy command threads. [fschulze]
Add option to log debug output. [fschulze]
Add helpers to setup proxycommand in plugins. [fschulze]

1.0.1 - 2014-08-13¶

Fix error output for plain instances on ssh connection failures. [fschulze]

1.0.0 - 2014-07-19¶

Fix removal of bad host keys when using non standard ssh port. [fschulze]
Renamed plain-master to plain, so the uids of instances are nicer. [fschulze]

1.0rc15 - 2014-07-16¶

Only remove bad host key from known_hosts instead of clearing it completely. [fschulze]
Removed support for proxyhost option. It caused hangs and failures on missing or invalid ssh fingerprints. [fschulze]
Allow empty startup_script option to mean use no startup script. [fschulze]

1.0rc14 - 2014-07-15¶

Allow fingerprint to be set to a public host key file. [fschulze]

1.0rc13 - 2014-07-08¶

Better error message for instances missing because the plugin isn’t installed. [fschulze]
Fix tests when ploy itself isn’t installed. [fschulze]

1.0rc12 - 2014-07-08¶

Use plain conftest.py instead of pytest plugin. [fschulze]

1.0rc11 - 2014-07-05¶

Fix uid method for master instances. [fschulze]

1.0rc10 - 2014-07-04¶

Print plugin versions with -v and --versions. [fschulze]
Python 3 compatibility. [fschulze]

1.0rc9 - 2014-06-29¶

Let plugins add type of lists to show with the list command. [fschulze]
Use server and instance consistently. [fschulze]
Always make instances accessible by their full name in the form of “[master_id]-[instance_id]”. Only if there is no conflict, the short version with just “[instance_id]” is also available for convenience. [fschulze]
Add instance id validator which limits to letters, numbers, dashes and underscores. [fschulze]
Renamed from mr.awsome to ploy. [fschulze]

1.0rc8 - 2014-06-16¶

Give a bit more info on ssh connection failures. [fschulze]

1.0rc7 - 2014-06-11¶

Expose some test fixtures for reuse in plugins. [fschulze]
Add before_terminate and after_start hooks and make it simple for plugins to add their own hooks. [fschulze]

1.0rc6 - 2014-06-10¶

Add get_path method to ConfigSection class. [fschulze]

1.0rc5 - 2014-06-09¶

Provide helper method ssh_args_from_info on BaseInstance to get the arguments for running the ssh executable from the info provided by init_ssh_key. [fschulze]
Allow overwriting the command name in help messages for bsdploy. [fschulze]
Make debug command usable for instances that don’t have a startup script. [fschulze]
Instances can provide a get_port method to return a default port. [fschulze]
Catch socket errors in init_ssh_key of plain instances to print additional info for debugging. [fschulze]
Delay setting of config file path to expose too early use of config in plugins. Refs #29 [fschulze]

1.0rc4 - 2014-05-21¶

Fix massagers for [instance:...] sections. [fschulze]
Copy massagers in ConfigSection.copy, so overrides in startup script work correctly. [fschulze]

1.0rc3 - 2014-05-15¶

Fetch fingerprints only when necessary. This speeds up connections when the fingerprint in known_hosts is still valid. [fschulze]

1.0rc2 - 2014-05-14¶

Moved setuptools-git from setup.py to .travis.yml, it’s only needed for releases and testing. [fschulze]
More tests. [fschulze]

1.0rc1 - 2014-03-23¶

Test, enhance and document adding massagers via config. [fschulze]
Moved ec2 and fabric integration into separate plugins. [fschulze]
You can now have instances with the same name if the belong to different masters, they will then get the name of the master as a prefix to their name. [fschulze]
Add possibility to overwrite the default config name. [tomster]
Improved proxycommand and documented it. [fschulze]
Make the AWS instance available in masters. This changes the get_masters plugin interface. [fschulze]
Use os.execvp instead of subprocess.call. This allows the use of assh in the proxycommand option, which greatly simplifies it’s use. [fschulze]
Added command plugin hooks. [fschulze]
The variable substitution for the proxycommand option now makes the other instances available in a dict under instances. And adds known_hosts. [fschulze]
Load plugins via entry points instead of the plugin section in the config. [fschulze]
Allow fallback to password for ssh to plain instances. [fschulze]
Add option to ask for manual fingerprint validation for plain instances. [fschulze]

0.13 - 2013-09-20¶

Use os.path.expanduser on all paths, so that one can use ~ in config values like the aws keys. [fschulze]

0.12 - 2013-09-11¶

There is no need to add the AWS account id to security group names anymore. [fschulze]
Rules are removed from security groups if they aren’t defined in the config. [fschulze]
Allow adding of custom config massagers from inside the config. [fschulze]
Support block device maps to enable use of more than one ephemeral disk. [fschulze]
Added do method on ec2 and plain instances which allows to call fabric commands. [fschulze]
Use PathMassager for access-key-id and secret-access-key in the ec2-master section. This might break existing relative paths for these options. [fschulze]
Added support for EBS boot instances. [fschulze]
Add option ssh-key-filename to point to a private ssh key for ec2 and plain instances. [fschulze]
Fix Fabric integration for newer versions of Fabric. [fschulze]
Support proxycommand option for plain instances. This also caused a change in the init_ssh_key API for plugins. [fschulze]
Support ProxyCommand from ~/.ssh/config for plain instances. Requires Fabric 1.5.0 and Paramiko 1.9.0 or newer. [fschulze]

0.11 - 2012-11-08¶

Support both the ssh and paramiko libraries depending on which Fabric version is used. [fschulze]

0.10 - 2012-06-04¶

Added ec2-connection which helps in writing Fabric scripts which don’t connect to a server but need access to the config and AWS (like uploading something to S3). [fschulze]
Fix several problems with using a user name other than root for the do and ssh commands. [fschulze]
Require Fabric >= 1.3.0. [fschulze]
Require boto >= 2.0. [fschulze]
Added hook for startup script options. [fschulze]
Added possibility to configure hooks. [fschulze]
Refactored to enable plugins for different virtualization or cloud providers. [fschulze]
Added lots of tests. [fschulze]

0.9 - 2010-12-09¶

Overwrites now also affect server creation, not just the startup script. [fschulze]
Added list command which supports just listing snapshots for now. [fschulze]
Added delete-volumes-on-terminate option to delete volumes created from snapshots on instance termination. [fschulze]
Added support for creating volumes from snapshots on instance start. [natea, fschulze]
Added support for ~/.ssh/config. This is a bit limited, because the paramiko config parser isn’t very good. [fschulze]
Added help command which provides some info for zsh autocompletion. [fschulze]

0.8 - 2010-04-21¶

For the do command the Fabric options reject_unknown_hosts and disable_known_hosts now default to true. [fschulze]
Allow adding normal servers to use with ssh and do commands. [fschulze]
Refactored ssh connection handling to only open network connections when needed. Any fabric option which doesn’t need a connection runs right away now (like -h and -l). [fschulze]
Fix status output after start. [fschulze]

0.7 - 2010-03-22¶

Added snapshot method to Server class for easy access from fabfiles. [fschulze]

0.6 - 2010-03-18¶

It’s now possible to specify files which contain the aws keys in the [aws] section with the access-key-id and secret-access-key options. [fschulze]
Added -c/--config option to specify the config file to use. [fschulze]
Added -v/--version option. [tomster (Tom Lazar), fschulze]
Comment lines in the startup script are now removed before any variables in it are expanded, not afterwards. [fschulze]
Use argparse library instead of optparse for more powerful command line parsing. [fschulze]

0.5 - 2010-03-11¶

Added gzipping of startup script by looking for gzip: prefix in the filename. [fschulze]
Added macro expansion similar to zc.buildout 1.4. [fschulze]

0.4 - 2010-02-18¶

Check console output in status and tell user about it. [fschulze]
Friendly message instead of traceback when trying to ssh into an unavailable server. [fschulze]
Remove comment lines from startup script if it’s starting with #!/bin/sh or #!/bin/bash. [fschulze]
Removed -r option for start and debug commands and replaced it with more general -o option. [fschulze]
Made startup script optional (not all AMIs support it, especially Windows ones). [fschulze]
The stop command actually only stops an instance now (only works with instances booted from an EBS volume) and the new terminate command now does what stop did before. [fschulze]
Better error message when no console output is available for ssh finger print validation. [fschulze]
Fixed indentation in documentation. [natea (Nate Aune), fschulze]

0.3 - 2010-02-08¶

Removed the [host_string] prefix of the do command output. [fschulze]

0.2 - 2010-02-02¶

Snapshots automatically get a description with date and volume id. [fschulze]
The ssh command can now be used with scp and rsync. [fschulze]

0.1 - 2010-01-21¶

Initial release [fschulze]

Plugins¶

ploy_ansible¶

Overview¶

The ploy_ansible plugin provides integration of Ansible with ploy. It automatically builds an inventory and provides a custom connection plugin.

Installation¶

ploy_ansible is best installed with easy_install, pip or with zc.recipe.egg in a buildout.

Commands¶

The plugin adds the following commands to ploy.

configure: Configures an instance. There are three ways to specify how to configure an instance. Applying the roles given by the roles option of an instance, a playbook set by the playbook option or a playbook with the unique name of the instance found in the playbooks-directory. Using roles or a playbook is mutually exclusive. If you specify a playbook and there is also a playbook in the default location, you will get a warning.
inventory: Lists all known groups and their associated hosts, including regular default groups, such as all but also implicit, ploy_ansible groups such as instances of a particular master (i.e. all ez-instances of an ez-master)
ansible: Runs an Ansible command. This basically reflects the ansible script of Ansible.
playbook: Applies a playbook. This basically reflects the ansible-playbook script of Ansible.
vault: Manages file encryption. This basically reflects the ansible-vault script of Ansible, but handles the encryption key source via ploy.conf.
vault-key: Manages the vault key.

Options¶

Global¶

playbooks-directory¶

The playbooks-directory option of the ansible section allows you to specify the directory where playbooks, roles, host_vars etc are looked up. If you specify a relative path, then it’s always relative to the ploy.conf directory. If you have a structure like this:

project
|-- deployment
| |-- roles
| |-- host_vars
|
|-- etc
  |-- ploy.conf

Then you would put the following into your ploy.conf:

[ansible]
playbooks-directory = ../deployment

By default it is set to the parent directory of the directory the ploy.conf is located at like this:

project
|-- roles
|-- host_vars
|-- etc
  |-- ploy.conf

vault-password-source¶

Using the keyring library, you can store the encryption key for the Ansible vault in your keychain.

The vault-password-source option is the id used in your keychain. The id must be unique among all people who have to use the feature, as it is used as an identifier in their keychain. If in doubt, use a speaking prefix and add a guid by running python -c "import uuid; print(uuid.uuid4().hex)".

If you want to rekey your files, you have to put the old id into the vault-password-old-source option and set a new id in vault-password-source. Just incrementing a number or appending a new guid is best.

Example:

[ansible]
vault-password-old-source = my-domain-deployment-0da2c8296f744c90a236721486dbd258
vault-password-source = my-domain-deployment-042a98b666ec4e4e8e06de7d42688f3b

You can manage your key with the vault-key command. For easy exchange with other developers, you can also export and import the key via gpg using the vault-key export and vault-key import commands.

Per instance¶

groups: Whitespace separated list of Ansible group names this instance should be added to in addition to the default ones.
roles: Used by the configure command. This allows you to configure an instance by applying the whitespace separated roles. This is like creating a playbook which only specifies a host and a list of roles names. If the sudo option is set, it’s also set for the generated playbook.
playbook: Allows you to explicitly specify a playbook to use for this instance. If you need sudo, then you have to add it yourself in that playbook.

Any option starting with ansible_ is passed through to Ansible as is. This can be used for settings like ansible_python_interpreter.

Any option starting with ansible- is stripped of the ansible- prefix and then passed through to Ansible. This is the main way to set Ansible variables for use in playbooks and roles.

All other options are prefixed with ploy_ and made available to Ansible.

Ansible inventory¶

All instances in ploy.conf are available to Ansible via their unique id.

The variables for each instance are gathered from group_vars, host_vars and the ploy.conf.

Ansible lookup plugins¶

The ploy_crypted lookup plugin can be used in playbooks to read the content of encrypted files. This is another way to access encrypted data where you don’t have to move that data into yml files. An added benefit is, that the file is only decrypted when it is actually accessed. If you run tasks filtered by tags and those tasks don’t access the encrypted data, then it’s not decrypted at all.

Warning

This lookup plugin only works with files that are plain ascii or utf-8. It’s a limitation caused by the way ansible handles variable substitution.

API usage¶

On the Python side, each ploy instance gains the following methods:

apply_playbook(self, playbook, *args, **kwargs): Applies the playbook to the instance.
has_playbook: Return True if the instance has either of the roles or a playbook option set.
get_playbook(*args, **kwargs): Returns an instance of the Ansible internal PlayBook class. This is either from a file (from playbook option or the playbook kwarg), or dynamically generated from the roles option.
configure(*args, **kwargs): Configures the instance with the same semantics as the configure command.
get_ansible_variables: Returns the Ansible variables from the inventory. This does not include facts, as it doesn’t connect to the instance. This is particularly useful in Fabric scripts.
get_vault_lib: Returns a readily usable Ansible VaultLib class. Use the encrypt and decrypt methods do encrypt/decrypt strings.

Changelog¶

1.4.0 - 2017-12-17¶

Look for [instance-name].yml in addition to [master-name]-[instance-name].yml. This allows using the same playbook for the same instance on multiple masters. [fschulze]

1.3.2 - 2016-06-02¶

Don’t add empty search path when no additional role or library paths are defined. This prevents the current working directory from being searched. [fschulze]

1.3.1 - 2015-09-03¶

Update Ansible requirement to < 2.dev0. The upcoming 2.0.0 has way too many internal changes to be supported. [fschulze]
Add hosts only once in Inventory. [fschulze]

1.3.0 - 2015-04-10¶

Added handling of groups option of instances to allow definition of additional Ansible groups. [fschulze]
Get host variables on demand instead of at startup. If you have many hosts with encrypted yml files, this speeds things up considerably in most cases. [fschulze]
Fixes for changes in ansible 1.9. [fschulze]
Added inventory command to list all known groups and their associated hosts. [fschulze]

1.2.4 - 2015-02-28¶

Pass on the sudo setting if the roles option is used. [fschulze]

1.2.3 - 2015-02-28¶

Fix sudo support for ansible > 1.6. [fschulze]
Print warning when using an untested version of ansible. [fschulze]
If ansible isn’t installed, then require >= 1.8 as that doesn’t violate the sandbox of buildout anymore. [fschulze]

1.2.2 - 2015-02-18¶

Test and fixes for changes in ansible 1.8. [fschulze]

1.2.1 - 2015-01-06¶

Limit Ansible to pre 1.8, as > 1.8 breaks stuff. [fschulze]

1.2.0 - 2014-10-27¶

Always set ansible_ssh_user in inventory. [fschulze]
Clear host and pattern cache after calling original Inventory.__init__ method. [fschulze]
Add --extra-vars option to configure command. [witsch (Andreas Zeidler)]
Provide ploy_crypted lookup plugin to load encrypted files into Ansible variables. Only ascii and utf8 encoded files will work. [fschulze]
Expand Ansible variables in get_ansible_variables method. [fschulze]
Support Ansible vault with safe key storage via keyring library, so you don’t have to type it in or have it in an unprotected file. [fschulze]

1.1.0 - 2014-08-13¶

Test and fixes for changes in ansible 1.7. [fschulze]
Add verbosity argument to configure command. [fschulze]

1.0.0 - 2014-07-19¶

Added documentation. [fschulze]

1.0b8 - 2014-07-15¶

Add ansible as dependency if it can’t be imported already. [fschulze]

1.0b7 - 2014-07-08¶

Packaging and test fixes. [fschulze]

1.0b6 - 2014-07-04¶

Use unique instance id to avoid issues. [fschulze]
Renamed mr.awsome to ploy and mr.awsome.ansible to ploy_ansible. [fschulze]

1.0b5 - 2014-06-16¶

Set user in playbook to the one from the config if it’s not set already. [fschulze]
Change default playbook directory from the aws.conf directory to it’s parent. [fschulze]

1.0b4 - 2014-06-11¶

Added playbook and roles config options for instances. [fschulze]
Added has_playbook and configure methods to instances. [fschulze]
Added before/after_ansible_configure hooks. [fschulze]

1.0b3 - 2014-06-09¶

Use execnet for connections. There is only one ssh connection per host and it’s reused for all commands. [fschulze]
Make sure the playbook directory is always absolute. [fschulze]
Prevent use of persistent ssh connections, as that easily results in connections to wrong jails because of the proxying. This makes ansible a lot slower at the moment. [fschulze]
Add support for su and vault (ansible 1.5) as well as --force-handlers (ansible 1.6). [fschulze]
Removed ansible from install requirements. It won’t install in a buildout so it needs to be installed in a virtualenv or via a system package. [fschulze]

1.0b2 - 2014-05-15¶

Add configure command which is a stripped down variant of the playbook command with assumptions about the location of the yml file. [fschulze]
Warn if a playbook is requested for a host that is not configured in the playbook hosts list. [fschulze]
Allow mr.awsome plugins to add ansible variables. [fschulze]
Inject the ansible paths sooner as they may not apply in some cases otherwise. [fschulze]
Moved setuptools-git from setup.py to .travis.yml, it’s only needed for releases and testing. [fschulze]

1.0b1 - 2014-03-24¶

Initial release [fschulze]

ploy_ec2¶

Overview¶

The ploy_ec2 plugin provides integration of Amazon EC2 with ploy.

Installation¶

ploy_ec2 is best installed with easy_install, pip or with zc.recipe.egg in a buildout.

Masters¶

To use ploy_ec2 you need an Amazon account and AWS keys.

Once you got your keys, you should put them in a secure location and reference them in your ploy.conf. Additionally you need to set the region of the master:

[ec2-master:ec2eu]
access-key-id = ~/.aws/ec2.id
secret-access-key = ~/.aws/ec2.key
region = eu-west-1

You can also set the AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY environment variables instead.

You need to define a master for each region you want to use.

Instances¶

Each instance has the following mandatory settings:

image: The Amazon Machine Image (AMI) that this instance will start up with.
keypair: The name of the SSH keypair to use.
placement: The availability zone in which to launch the instances.
securitygroups: The name of the Securitygroups this instance should be assigned to.

The following settings are optional:

instance_type

ip

startup_script: Path to a script which will be run right after creation and first start of the instance. This uses the User Data feature and needs to be supported by the AMI.

volumes

snapshots

device_map

delete-volumes-on-terminate

Securitygroups¶

description

connections

[ec2-securitygroup:app-server]
description = The production server
connections =
    tcp     22      22      0.0.0.0/0
    tcp     80      80      0.0.0.0/0

Volumes¶

You can define volumes via ec2-volume sections. The id of the section must not start with vol-. You can declare the size as a number of GB.

If the volume doesn’t exist, it is automatically created.

[ec2-volume:a-volume-name]
size = 100

[ec2-instance:foo]
...
volumes = a-volume-name /dev/sdf

Macro expansion¶

For instances the ip and volumes options aren’t copied when expanding macros.

Fingerprint verification¶

Automatic ssh fingerprint verification works by checking whether the fingerprint is in the console output of the instance.

After reboot or stop/start of an instance, the console output is refreshed. The problem with that is, that the fingerprint isn’t included in the console anymore by default. To fix that you need to log the fingerprint on reboot somehow. One way to do that with Ubuntu is to add a script at /var/lib/cloud/scripts/per-boot/ssh-keys with this content:

#!/bin/sh
/usr/bin/ssh-keygen -l -f /etc/ssh/ssh_host_rsa_key.pub

Make sure that script is executable.

Changelog¶

1.2.1 - Unreleased¶

Support yet another ssh key output format from console output of instance. [fschulze]

1.2.0 - 2015-09-03¶

Check status of volume to give helpful error message if it’s still attached. [fschulze]
Allow volume definition via ec2-volume sections. [fschulze]
Fix support of console output for the ploy debug -c command. [fschulze]
Reuse init_ssh_key from ploy.plain to get way more options and error checking. [fschulze]

1.1.1 - 2015-01-22¶

Only set device_map if it’s in the config, the previous None default didn’t always work. [fschulze]
Fixed console output availability test for the status command. [fschulze]
Better error message if fingerprint isn’t in console output. [fschulze]
There can be multiple instances for the same name if they were quickly started and stopped. Handle that case when requesting status of master. [fschulze]

1.1.0 - 2014-10-27¶

Print status of all ec2 instances when requesting status of master. [fschulze]

1.0.0 - 2014-07-19¶

Added documentation. [fschulze]

1.0b4 - 2014-07-15¶

Fix confusion between instance from ploy and ec2 instance. [fschulze]

1.0b3 - 2014-07-08¶

Moved snapshots list command here after ploy enabled it. [fschulze]
Renamed mr.awsome to ploy and mr.awsome.ec2 to ploy_ec2. [fschulze]

1.0b2 - 2014-05-15¶

Renamed conn to ec2_conn to allow reuse of conn from BaseInstance. [fschulze]
Moved setuptools-git from setup.py to .travis.yml, it’s only needed for releases and testing. [fschulze]

1.0b1 - 2014-03-24¶

Initial release [fschulze]

ploy_ezjail¶

Overview¶

The ploy_ezjail plugin provides integration of ezjail with ploy to manage FreeBSD jails.

Installation¶

ploy_ezjail is best installed with easy_install, pip or with zc.recipe.egg in a buildout.

Masters¶

To use ploy_ezjail you need a host running FreeBSD on which you want to manage jails.

You declare a master with [ez-master:masterid] where masterid is the name you want to use for this master. Now you can either add options like for a plain ploy instance, or you can use the instance option to refer to another instance from your config like this:

[ez-master:master1]
host = myhost.example.com

[plain-instance:foohost]
host = foohost.example.com

[ez-master:master2]
instance = foohost

The latter is most useful in conjunction with other ploy backend plugins, as it allows you to easily switch between provisioners, i.e. to have an ez-master provisioned on VirtualBox during development and on a plain instance in production.

Options¶

debug-commands: If set to yes, the commands executed on the host are echoed locally.
instance: The instance to use as host for this master. If empty, the local machine is used without an ssh connection.
ezjail-admin: Path to the ezjail-admin script on the host. Defaults to /usr/local/bin/ezjail-admin.
sudo: Use sudo to run commands on the host.

Instances¶

At the moment all jails will be created using ZFS (the -c zfs option of ezjail-admin), so the host needs to be setup accordingly.

Options¶

ip

The ip address to use for the jail.

This can either be a single IPv4 address:

ip = 10.0.0.3

or any number of IPv4 and IPv6 addresses attached to different devices:

ip = lo1|10.0.0.3,vtnet0|2a03:b0c0:3:d0::3a4d:c002

The latter format is ezjail’s own. Required

flavour

The flavour to use for this jail. This is explained in the ezjail docs.

ezjail-name

The name to use for the jail. By default the id of the instance is used.

mounts

Additional mount points for the jail. You can specify one mount point per line. The format is:

src=SRC dst=DST [ro=true] [create=true]

The src is the path on the host, dst is the path inside the jail.

If ro is set to true, then the mount is read only.

When create is enabled, then the src path is created with mkdir -p. The dst path is always created inside the jail with mkdir -p.

You can reference ZFS sections inside src with {zfs[name]} where name is the ez-zfs section name. You can use the name of the jail instance with {name}} in both src and dst. Examples:

src=/foo dst=/foo
src={zfs[backup]} dst=/bak
src={zfs[data]}/{name} dst=/mnt/data create=true
src={zfs[static]} dst=/mnt/static ro=true

no-terminate

If set to yes, the jail can’t be terminated via ploy until the setting is changed to no or removed entirely.

startup_script

Path to a local script (relative to the location of the configuration file) which will be run inside the jail right after creation and first start of the jail.

rc_require

String that indicates which other jails this jail requires to start up, effectively allowing you to define the startup order of jails. See rcorder(8) for more details. This value is written upon each startup of the jail not just when it is created initially, so to have changes take effect, it’s sufficient to restart it. Optional

rc_provide

String that indicates what this jail provides. ezjail itself always sets its jails to provide standard_ezjail to which ploy_ezjail adds the name of the jail. IOW if you simply want to build a startup order using the names of the jails, you will not need to set this value. If you want this jail to provide any additional values, set them here. This value is written upon each startup of the jail not just when it is created initially, so to have changes take effect, it’s sufficient to restart it. Optional

ZFS sections¶

You can specify ZFS filesystems via [ez-zfs:name] sections. This is used in mounts of jails to get the mountpoint and verify that the path exists and is it’s own ZFS filesystem. You can also create new ZFS filesystems with the create option.

Options¶

create

If set to yes, the filesystem is created when first used.

path

Specifies the path of this filesystem. This is not the mountpoint, but the ZFS path. You can reference other ZFS sections with {zfs[name][path]}. The name is the name of the referenced ZFS section. The [path] at the end is mandatory, as otherwise you would get the mountpoint of the referenced ZFS section. Examples:

[ez-zfs:data]
path = tank/data

[ez-zfs:shared]
path = {zfs[data][path]}/shared

[ez-zfs:jails]
path = {zfs[data][path]}/jails

[ez-zfs:backup]
create = true
path = tank/backup

Changelog¶

1.5.2 - Unreleased¶

1.5.1 - 2018-02-03¶

Fix startup script for FreeBSD 11.x by replacing dots in it’s name with underscores. [fschulze]

1.5.0 - 2017-12-17¶

Add get_fingerprints to support host keys with all key types. [fschulze]
Fix jail host status command for various cases. [fschulze]
Output stdout in addition to stderr on various errors. This let’s one see additional debug info when jails don’t start. [fschulze]

1.4.0 - 2015-10-16¶

Allow setting the jail name via ezjail-name instead of using the instance id as the default. [fschulze]

1.3.0 - 2015-09-03¶

Improved error handling with useful error messages instead of tracebacks. [fschulze]
Allow setting startup order of jails. [tomster]

1.2.0 - 2015-03-05¶

Use new Executor helper from ploy 1.2.0 which handles ssh agent forwarding. [fschulze]
Enable “local mode” where if the instance option is empty all commands are executed locally. [fschulze]

1.1.0 - 2014-10-27¶

Print status of all jails when requesting status of master. [fschulze]
Check jail status before trying to connect. [fschulze]
Use new helper in ploy 1.0.2 to setup proxycommand. [fschulze]

1.0.0 - 2014-07-19¶

Added documentation. [fschulze]

1.0b9 - 2014-07-08¶

Packaging and test fixes. [fschulze]

1.0b8 - 2014-07-04¶

Python 3 compatibility. [fschulze]
Renamed mr.awsome to ploy and mr.awsome.ezjail to ploy_ezjail. [fschulze]

1.0b7 - 2014-06-16¶

Provide default values for proxyhost and proxycommand options. [fschulze]
Merge config of ez-master with the instance it’s using. [fschulze]

1.0b6 - 2014-06-11¶

Pass changes of proxy instance config on to the proxied instance config. [fschulze]

1.0b5 - 2014-06-10¶

Forcefully destroy jail. Together with ezjail 3.4.1 this solves the issue that sometimes the ZFS filesystem wasn’t removed and the jail couldn’t be started without manual intervention. [fschulze]

1.0b4 - 2014-05-22¶

Clear out massagers after copying the config for the proxy instance to prevent conflicts when the proxy instance is created. [fschulze]

1.0b3 - 2014-05-21¶

Fixes to make [instance:...] using an ez-master work. [fschulze]

1.0b2 - 2014-05-15¶

Added instance option to ez-master section to use another instance as the jail host. [fschulze, tomster]
Moved setuptools-git from setup.py to .travis.yml, it’s only needed for releases and testing. [fschulze]

1.0b1 - 2014-03-24¶

Initial release [fschulze]

ploy_fabric¶

Overview¶

The ploy_fabric plugin provides integration of Fabric with ploy.

Installation¶

ploy_fabric is best installed with easy_install, pip or with zc.recipe.egg in a buildout.

Once installed, it’s functionality is immediately usable with ploy.

Commands¶

The plugin adds the following commands to ploy.

do

Runs a Fabric task with simplified syntax for arguments. You can just put positional arguments on the command line behind the task name. For keyword arguments user the name=value syntax. For example:

ploy do something arg key=value

fab

Runs a Fabric task and passes on command line options to Fabric. This basically reflects the fab script of Fabric.

Options¶

Instances only get the new fabfile option to specify which file to look in for tasks. The location is relative to ploy.conf.

Instance methods¶

For the Python side, each instance gains the do(task, *args, **kwargs) method. The task argument is the name of a task from the Fabric script which should be run. The remaining arguments are passed on to that task.

Another helper added to each instance is a context manager accessible via the fabric attribute on instances. With that you can switch to a new ssh connection with a different user in your Fabric tasks:

from fabric.api import env, run


def sometask():
    run("whoami")  # prints the default user (root)
    with env.instance.fabric(user='foo'):
      run("whoami")  # prints 'foo' if the connection worked
    run("whoami")  # prints the default user (root)

All changes to the Fabric environment are reverted when the context manager exits.

Fabric task decorator¶

With ploy_fabric.context you can decorate a task to use a specific user with a separate connection. All changes to the Fabric environment are reverted when the context manager exits. This is useful if you want to run a task from inside another task.

from fabric.api import env, run
from ploy_fabric import context


@context  # always run with the default user
def sometask():
    run("whoami")  # prints the default user (root)


@context(user=None)  # always run with the default user (alternate syntax)
def someothertask():
    env.forward_agent = True
    run("whoami")  # prints the default user (root)


@context(user='foo')  # always run as foo user
def anothertask():
    env.forward_agent = False
    run("whoami")  # prints the default user (user)
    someothertask()
    assert env.forward_agent == False

Fabric environment¶

The Fabric environment has the following settings by default.

reject_unknown_hosts: Always set to True, ssh connections are handled by this plugin and ploy.
disable_known_hosts: Always set to True, handled by ploy.
host_string: The unique id of the current instance, only manipulate if you know what you do!
known_hosts: Path to the known_hosts file managed by ploy.
instances: Dictionary allowing access to the other instances to get variables or call methods on.
instance: The current instance to access variables from the config attribute or other things and methods.
config_base: The directory of ploy.conf.

Any option of the instance starting with fabric- is stripped of the fabric- prefix and overwrites settings in the environment with that name.

Changelog¶

1.1.1 - Unreleased¶

1.1.0 - 2014-10-27¶

Require Fabric >= 1.4.0 and vastly simplify the necessary patching. [fschulze]
Close all newly opened connections after a Fabric call. [fschulze]
Add context manager and decorator to easily switch fabric connections. [fschulze]

1.0.0 - 2014-07-19¶

Added documentation. [fschulze]

1.0b6 - 2014-07-15¶

Allow overwriting of fabric env from config with options prefixed by fabric-, i.e. fabric-shell = /bin/sh -c. [fschulze]

1.0b5 - 2014-07-08¶

Packaging and test fixes. [fschulze]
Fix task listing for do command. [fschulze]

1.0b4 - 2014-07-04¶

Use unique id for host string to avoid issues. [fschulze]
Added fab command which is just a wrapper for Fabric with all it’s options and reworked do command into a simple version to just run a task. [fschulze]
Renamed mr.awsome to ploy and mr.awsome.fabric to ploy_fabric. [fschulze]

1.0b3 - 2014-06-09¶

When depending on Fabric, skip 1.8.3 which added a version pin on paramiko. [fschulze]
Only add Fabric to install_requires if it can’t be imported. That way we don’t get problems if it’s already installed as a system packages or in a virtualenv. [fschulze]

1.0b2 - 2014-05-15¶

Register fabfile massager for all instances. [fschulze]
Use context manager for output filtering and filter in do helper. [fschulze]
Moved setuptools-git from setup.py to .travis.yml, it’s only needed for releases and testing. [fschulze]

1.0b1 - 2014-03-24¶

Initial release [fschulze]

ploy_openvz¶

Changelog¶

1.0b3 - unreleased¶

Renamed mr.awsome to ploy and mr.awsome.openvz to ploy_openvz. [fschulze]

1.0b2 - 2014-05-15¶

Moved setuptools-git from setup.py to .travis.yml, it’s only needed for releases and testing. [fschulze]

1.0b1 - 2014-03-24¶

Initial release [fschulze]

ploy_virtualbox¶

Overview¶

The ploy_virtualbox plugin provides integration of VirtualBox with ploy.

Installation¶

ploy_virtualbox is best installed with easy_install, pip or with zc.recipe.egg in a buildout.

Master¶

The default master for ploy_virtualbox is virtualbox and has the following options:

headless: Whether to start instances in headless mode by default. If not set, the system default is used.
use-acpi-powerbutton: Whether to use acpi power off by default when stopping instances. If not set, the system default is used.
basefolder: The basefolder for VirtualBox data. If not set, the VirtualBox default is used. This varies depending on the OS that ploy is running in. When not provided as absolute path, then it’s relative to ploy.conf.
instance: Name of instance to use to execute VirtualBox commands instead of the default local machine.

Example:

[vb-master:virtualbox]
headless = true

Instances¶

headless: Whether to start this instance in headless mode. If not set, the setting of the master is used.
use-acpi-powerbutton: Whether to use acpi power off for this instance when stopping. If not set, the setting of the master is used.
basefolder: The basefolder for this instances VirtualBox data. If not set, the setting of the master is used. When not provided as absolute path, then it’s relative to ploy.conf.
no-terminate: If set to yes, the instance can’t be terminated via ploy until the setting is changed to no or removed entirely.

Any option starting with vm- is stripped of the vm- prefix and passed on to VBoxManage. Almost all of these options are passed as is. The following options are handled differently or have some convenience added:

storage

One storage definition per line. See VBoxManage storageattach documentation for details.

If you don’t specify the type, then hdd is used by default.

You don’t have to specify the port, if not set the line number starting at zero is used.

If you don’t set --storagectl then sata is used as default and if that controller doesn’t exist it’s created automatically.

If medium references a local file that path will be passed directly to VBoxManage storageattach.

If it takes the form vb-disk:NAME which refers to a Disk section called NAME that will be used instead.

If it takes the form of an URL, the filename of that URL is assumed to be located at ~/.ploy/downloads/ (this default can be overridden in the [global] section of the configuration file with an entry download_dir). If the file does not exist it will be downloaded.

When using the URL notation it is strongly encouraged to also provide a checksum using the --medium_sha1 key (currently only SHA1 is supported).

Example for using a local ISO image as DVD drive:

storage =
    --type dvddrive --medium ~/downloads/archives/mfsbsd-se-9.2-RELEASE-amd64.iso
    --medium vb-disk:boot

Example for referencing an external URL:

storage =
    --type dvddrive --medium http://mfsbsd.vx.sk/files/iso/11/amd64/mfsbsd-se-11.1-RELEASE-amd64.iso --medium_sha1 2fd80caf57c29d62859bccfa3b8ec7b5b244406e
    --medium vb-disk:boot

hostonlyadapter

If there is a matching Host only interface section, then that is evaluated.

Disk sections¶

These section allow you to describe how disks should be created.

You can set the size, variant and format options as described in the VBoxManage createhd documentation.

The filename option allows you to set a filename for the disk, the extension is automatically added based on the format option. If you use a relative path, then it’s base is the basefolder setting of the instance.

When the delete option is set to false, the disk is not deleted when the instance using it is terminated. The default is to delete the disk upon instance termination.

Example:

[vb-disk:boot]
size = 102400

Host only interface sections¶

If you want to use host only network interfaces, then this allows you to make sure your settings are as expected and the interface exists. For now only the ip option is supported. See VBoxManage hostonlyif documentation for details.

Example:

[vb-hostonlyif:vboxnet0]
ip = 192.168.56.1

DHCP¶

If a vb-dhcpserver section with the same name exists, then it is checked and if needed configured as well. See VBoxManage hostonlyif documentation for details.

Example:

[vb-dhcpserver:vboxnet0]
ip = 192.168.56.2
netmask = 255.255.255.0
lowerip = 192.168.56.100
upperip = 192.168.56.254

The combination of vb-hostonlyif with vb-dhcpserver allows to configure a hostonly network with a deterministic IP address. In the above example you could configure an instance with a static IP address of 192.168.56.99 which would be addressable from the host. The important part is to chose an address that is within the DHCP server network but outside its DHCP pool, which is defined by lowerip and upperip respecitively.

SSH¶

Depending on the setup we can’t get the IP address or host name automatically.

Unfortunately VirtualBox doesn’t provide a way to see which instance got which IP address from it’s own DHCP servers for example.

If you know which host name or ip address your instance will have, then set the host or ip option as explained above in the hostonly section.

As a workaround you can also setup a NAT port forwarding like this:

vm-nic2 = nat
vm-natpf2 = ssh,tcp,,47022,,22

For this case ploy_virtualbox knows how to get the port and uses it for SSH access via localhost.

If you install the VirtualBox guest additions in your instance, then the status command can show you the current IP address of the instance.

Example config¶

[vb-master:virtualbox]
# use-acpi-powerbutton = false

[vb-disk:boot]
size = 102400

[vb-hostonlyif:vboxnet0]
ip = 192.168.56.1

[vb-dhcpserver:vboxnet0]
ip = 192.168.56.2
netmask = 255.255.255.0
lowerip = 192.168.56.100
upperip = 192.168.56.254

[vb-instance:foo]
# headless = true
vm-ostype = FreeBSD_64
vm-memory = 512
vm-accelerate3d = off
vm-acpi = on
vm-rtcuseutc = on
vm-boot1 = disk
vm-boot2 = dvd
vm-nic1 = hostonly
vm-hostonlyadapter1 = vboxnet0
vm-nic2 = nat
vm-natpf2 = ssh,tcp,,47022,,22
storage =
    --type dvddrive --medium ~/downloads/archives/mfsbsd-se-9.2-RELEASE-amd64.iso
    --medium vb-disk:boot

Changelog¶

1.2.1 - Unreleased¶

1.2.0 - 2015-09-03¶

Add delete option to disks. If set to false, the disk is kept in place upon instance termination and not deleted as per default. [fschulze]
Ask user whether to continue or not when checksum of downloaded iso image doesn’t match. [tomster]

1.1.0 - 2015-01-20¶

Log info when starting an instance. [fschulze]
Handle instances in aborted state. [fschulze]
Print error output of commands on failures. [fschulze]
Use new helper in ploy 1.0.2 to setup proxycommand. [fschulze]
Added possibility to specify a remote instance to use for a virtualbox master. [fschulze]
Added ability to reference disk images via external URL for virtualbox instances storage. [tomster]

1.0.0 - 2014-07-19¶

Added documentation. [fschulze]
Renamed vb-master to virtualbox, so the uids of instances are nicer. [fschulze]
Enable DHCP server when creating or modifying it. [fschulze]

1.0b4 - 2014-07-15¶

Verify and if possible create host only interfaces and dhcpservers. [fschulze]
Add support for instances that have manually been put into saved state. [fschulze]

1.0b3 - 2014-07-08¶

Packaging and test fixes. [fschulze]

1.0b2 - 2014-07-04¶

Python 3 compatibility. [fschulze]
Renamed mr.awsome to ploy and mr.awsome.virtualbox to ploy_virtualbox. [fschulze]

1.0b1 - 2014-06-16¶

Initial release [fschulze]