Synopsis

Sparrowdo provision for Terraform backed instances.

Install

$ zef install Sparrowform

Limitations

Currently only ssh accessed instances with public IPs are supported ( aws ec2 / google compute instances with public IPs ).

Terrafrom resources supported:

• aws_instances
• google_compute_instance

Ping me if you need more Terraform resourses support.

Usage

Write some Terraform scenarios and deploy some instances

$ terraform apply

Create Sparrowodo scenarios, one per instance

Scenarios should be named as $terrafrom-instance-type.$terraform-instance-ID.sparrowfile

$ nano aws_instance.example.sparrowfile
$ nano aws_instance.example2.sparrowfile
$ nano aws_instance.example3.sparrowfile
# ...

See also Sparrowdo one liners option on how to run sparrowdo tasks/modules not scenarios.

Run Sparrowdo provision

This command will run Sparrowdo scenarios for all instances for which files $terrafrom-instance-ID.sparrowfile exist:

$ sparrowform

Handling ssh connections

You may pass ssh connection parameters by specifying sparrowdo cli parameters:

$ sparrowform --ssh_user=ec2-user --ssh_private_key=/path/to/ssh.key

Using sparrowdo one liners instead of scenarios:

# install Nginx on all instances:
$ sparrowform --module_run=Nginx

# check if Nginx alive on all instances:
$ sparrowform --task_run=bash@command='"ps uax|grep nginx"'

# install packages
$ sparrowform --task_run=package-generic@list="'nano mc'"

Default Sparrowdo scenario

If you don't want bother with creating scenarios for every instance, you may choose to defined default scenario.

Create scenario named sparrowfile:

$ nano sparrowfile

bash "apt-get update";

So, these instances which do not have a related Sparrowdo scenarios files will use this default scenario.

Terraform resources access

Sparrowform exposes nice API to access Terraform internal guts inside Sparrowdo scenarios.

The function tf-resources returns Perl6 Array of all Terraform resources. Each elements consists of two elements, the first one holds resource identificator, the second one holds resource data, represented as Perl6 Hash.

Here is usage example:

$ cat sparrowfile 

# let's insert all ec2 instances DNS names into ever instance's /etc/hosts file:

use Sparrowform;

my @hosts = (
  "127.0.0.1   localhost localhost.localdomain localhost4 localhost4.localdomain4",
  "::1         localhost localhost.localdomain localhost6 localhost6.localdomain6"
);

for tf-resources() -> $r {
  my $rd = $r[1]; # resource data
  next unless $rd<public_ip>;
  next unless $rd<public_dns>;
  next if $rd<public_ip> eq input_params('Host');
  push @hosts, $rd<public_ip> ~ ' ' ~ $rd<public_dns>;
}

file '/etc/hosts', %(
  action  => 'create',
  content => @hosts.join("\n")
);

Debugging

If something goes awry ... Enable SPF_DEBUG variable to see internal output:

$ SPF_DEBUG=1 sparrowform

Dry run

If you only want to see which instances would be deployed, run with SPG_DRYRUN enabled:

$ SPF_DRYRUN=1 sparrowform

Author

Alexey Melezhik

See also

• Sparrowdo
• Terraform