github.com/hobbeswalsh/terraform@v0.3.7-0.20150619183303-ad17cf55a0fa/website/source/intro/getting-started/dependencies.html.md (about) 1 --- 2 layout: "intro" 3 page_title: "Resource Dependencies" 4 sidebar_current: "gettingstarted-deps" 5 description: |- 6 In this page, we're going to introduce resource dependencies, where we'll not only see a configuration with multiple resources for the first time, but also scenarios where resource parameters use information from other resources. 7 --- 8 9 # Resource Dependencies 10 11 In this page, we're going to introduce resource dependencies, 12 where we'll not only see a configuration with multiple resources 13 for the first time, but also scenarios where resource parameters 14 use information from other resources. 15 16 Up to this point, our example has only contained a single resource. 17 Real infrastructure has a diverse set of resources and resource 18 types. Terraform configurations can contain multiple resources, 19 multiple resource types, and these types can even span multiple 20 providers. 21 22 On this page, we'll show a basic example of multiple resources 23 and how to reference the attributes of other resources to configure 24 subsequent resources. 25 26 ## Assigning an Elastic IP 27 28 We'll improve our configuration by assigning an elastic IP to 29 the EC2 instance we're managing. Modify your `example.tf` and 30 add the following: 31 32 ``` 33 resource "aws_eip" "ip" { 34 instance = "${aws_instance.example.id}" 35 } 36 ``` 37 38 This should look familiar from the earlier example of adding 39 an EC2 instance resource, except this time we're building 40 an "aws\_eip" resource type. This resource type allocates 41 and associates an 42 [elastic IP](http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/elastic-ip-addresses-eip.html) 43 to an EC2 instance. 44 45 The only parameter for 46 [aws\_eip](/docs/providers/aws/r/eip.html) is "instance" which 47 is the EC2 instance to assign the IP to. For this value, we 48 use an interpolation to use an attribute from the EC2 instance 49 we managed earlier. 50 51 The syntax for this interpolation should be straightforward: 52 it requests the "id" attribute from the "aws\_instance.example" 53 resource. 54 55 ## Plan and Execute 56 57 Run `terraform plan` to view the execution plan. The output 58 will look something like the following: 59 60 ``` 61 $ terraform plan 62 ... 63 64 + aws_eip.ip 65 instance: "" => "${aws_instance.example.id}" 66 private_ip: "" => "<computed>" 67 public_ip: "" => "<computed>" 68 69 + aws_instance.example 70 ami: "" => "ami-aa7ab6c2" 71 availability_zone: "" => "<computed>" 72 instance_type: "" => "t1.micro" 73 key_name: "" => "<computed>" 74 private_dns: "" => "<computed>" 75 private_ip: "" => "<computed>" 76 public_dns: "" => "<computed>" 77 public_ip: "" => "<computed>" 78 security_groups: "" => "<computed>" 79 subnet_id: "" => "<computed>" 80 ``` 81 82 Terraform will create two resources: the instance and the elastic 83 IP. In the "instance" value for the "aws\_eip", you can see the 84 raw interpolation is still present. This is because this variable 85 won't be known until the "aws\_instance" is created. It will be 86 replaced at apply-time. 87 88 Next, run `terraform apply`. The output will look similar to the 89 following: 90 91 ``` 92 aws_instance.example: Creating... 93 ami: "" => "ami-aa7ab6c2" 94 instance_type: "" => "t1.micro" 95 aws_eip.ip: Creating... 96 instance: "" => "i-0e737b25" 97 98 Apply complete! Resources: 2 added, 0 changed, 0 destroyed. 99 ``` 100 101 It is clearer to see from actually running Terraform, but 102 Terraform creates the EC2 instance before the elastic IP 103 address. Due to the interpolation earlier where the elastic 104 IP requires the ID of the EC2 instance, Terraform is able 105 to infer a dependency, and knows to create the instance 106 first. 107 108 ## Implicit and Explicit Dependencies 109 110 Most dependencies in Terraform are implicit: Terraform is able 111 to infer dependencies based on usage of attributes of other 112 resources. 113 114 Using this information, Terraform builds a graph of resources. 115 This tells Terraform not only in what order to create resources, 116 but also what resources can be created in parallel. In our example, 117 since the IP address depended on the EC2 instance, they could 118 not be created in parallel. 119 120 Implicit dependencies work well and are usually all you ever need. 121 However, you can also specify explicit dependencies with the 122 `depends_on` parameter which is available on any resource. For example, 123 we could modify the "aws\_eip" resource to the following, which 124 effectively does the same thing and is redundant: 125 126 ``` 127 resource "aws_eip" "ip" { 128 instance = "${aws_instance.example.id}" 129 depends_on = ["aws_instance.example"] 130 } 131 ``` 132 133 If you're ever unsure about the dependency chain that Terraform 134 is creating, you can use the [`terraform graph` command](/docs/commands/graph.html) to view 135 the graph. This command outputs a dot-formatted graph which can be 136 viewed with 137 [Graphviz](http://www.graphviz.org/). 138 139 ## Non-Dependent Resources 140 141 We can now augment the configuration with another EC2 instance. 142 Because this doesn't rely on any other resource, it can be 143 created in parallel to everything else. 144 145 ``` 146 resource "aws_instance" "another" { 147 ami = "ami-aa7ab6c2" 148 instance_type = "t1.micro" 149 } 150 ``` 151 152 You can view the graph with `terraform graph` to see that 153 nothing depends on this and that it will likely be created 154 in parallel. 155 156 Before moving on, remove this resource from your configuration 157 and `terraform apply` again to destroy it. We won't use the 158 second instance anymore in the getting started guide. 159 160 ## Next 161 162 In this page you were introduced to both multiple resources 163 as well as basic resource dependencies and resource attribute 164 interpolation. 165 166 Moving on, [we'll use provisioners](/intro/getting-started/provision.html) 167 to do some basic bootstrapping of our launched instance.