Please refer to below code
data "amazon-ami" "ubuntu" {
most_recent = true
filters = {
virtualization-type = "hvm"
name = "ubuntu/images/*ubuntu-focal-20.04-amd64-server-*"
root-device-type = "ebs"
}
owners = ["099720109477"]
}
source "null" "one" {
communicator = "none"
}
build {
name = "source-ouput"
sources = ["source.null.one"]
provisioner "shell-local" {
inline = [
"echo ${data.amazon-ami.ubuntu}", ##### <--- problem happens here
]
}
}
As you can see in the code, I want to print out the return value of data "amazon-ami" "ubuntu"
If I run this with packer build .. error occurs like below
Error: Failed preparing provisioner-block "shell-local" ""
on main.pkr.hcl line 20:
(source code not available)
main.pkr.hcl:22,15-37: Invalid template interpolation value; Cannot include the
given value in a string template: string required.
Is there any solution to print the object that returned by data?
Related
In a aws_ssoadmin_permission_set_inline_policy ressource, i'm using a for_each to parse a list of name corresponding to my data source name. It doesn't work when using the each.key but wokring when hard coding the value inline_policy = data.aws_iam_policy_document.emobg-sso-billing-admin.json
data "aws_iam_policy_document" "emobg-sso-billing-admin" {
statement {
sid = "VisualEditor0"
effect = "Allow"
actions = [
"aws-marketplace:*",
"aws-portal:*",
"budgets:*"
]
resources = [
"*",
]
}
}
data "aws_iam_policy_document" "emobg-sso-billing-audit" {
statement {
sid = "VisualEditor0"
effect = "Allow"
actions = [
"support:*",
"tag:*",
"s3:*"
]
resources = [
"*",
]
}
}
resource "aws_ssoadmin_permission_set" "emobg" {
for_each = toset(local.permission_sets_name)
name = each.key
description = each.key
instance_arn = local.sso_instance_arn
session_duration = local.session_duration
}
resource "aws_ssoadmin_permission_set_inline_policy" "emobg" {
for_each = toset(local.permission_sets_name)
inline_policy = format("data.aws_iam_policy_document.%s.json", each.key) # <-- doesn't works
# inline_policy = data.aws_iam_policy_document.emobg-sso-billing-admin.json # <-- works
instance_arn = local.sso_instance_arn
permission_set_arn = aws_ssoadmin_permission_set.emobg[each.key].arn
}
locals {
session_duration = "PT8H"
permission_sets_name = [
"emobg-sso-billing-admin",
"emobg-sso-billing-audit",
]
}
The error message is:
2022-11-01T01:19:43.923+0100 [ERROR] vertex "aws_ssoadmin_permission_set_inline_policy.emobg[\"emobg-sso-billing-admin\"]" error: "inline_policy" contains an invalid JSON policy
2022-11-01T01:19:43.923+0100 [ERROR] vertex "aws_ssoadmin_permission_set_inline_policy.emobg (expand)" error: "inline_policy" contains an invalid JSON policy
╷
│ Error: "inline_policy" contains an invalid JSON policy
│
│ with aws_ssoadmin_permission_set_inline_policy.emobg["emobg-sso-billing-admin"],
│ on permission_set.tf line 13, in resource "aws_ssoadmin_permission_set_inline_policy" "emobg":
│ 13: inline_policy = format("data.aws_iam_policy_document.%s.json", each.value)
I really don't understand what's wrong with the JSON policy because it's the same.
Maybe I missed something ?
Because you are using format("data.aws_iam_policy_document.%s.json", each.key), the policy will be literal string "data.aws_iam_policy_document.%s.json".
You have only single policy, so you have to use it directly:
inline_policy = data.aws_iam_policy_document.emobg-sso-billing-admin.json
that's why it works. You do not have more then one aws_iam_policy_document in your code.
Thanks to Marcin, who give me the anser in comment: You can't do what you want. TF does not support dynamic references to different resources.
As it was mentioned, it's not allowed from terraform to make a dynamic references, so I finally used a map even if the name of the policy is the same of the base name.
data "aws_iam_policy_document" "emobg-sso-billing-admin" {
statement {
sid = "VisualEditor0"
effect = "Allow"
actions = [
"aws-marketplace:*",
"aws-portal:*",
"budgets:*"
]
resources = [
"*",
]
}
}
data "aws_iam_policy_document" "emobg-sso-billing-audit" {
statement {
sid = "VisualEditor0"
effect = "Allow"
actions = [
"support:*",
"tag:*",
"s3:*"
]
resources = [
"*",
]
}
}
... [ ALL OTHERS DATA SOURCES POLICIES ARE LISTED HERE ]
resource "aws_ssoadmin_permission_set" "emobg" {
for_each = local.permission_set_map
name = each.key
description = each.key
instance_arn = local.sso_instance_arn
session_duration = local.session_duration
}
resource "aws_ssoadmin_permission_set_inline_policy" "emobg" {
for_each = local.inline_policies_map
inline_policy = each.value
instance_arn = local.sso_instance_arn
permission_set_arn = aws_ssoadmin_permission_set.emobg[each.key].arn
}
locals {
session_duration = "PT8H"
permission_set_map = { for ps in local.permission_sets : ps.name => ps }
inline_policies_map = { for ps in local.permission_sets : ps.name => ps.inline_policy if ps.inline_policy != "" }
}
locals {
permission_sets = [
{
name = "emobg-sso-billing-admin",
inline_policy = data.aws_iam_policy_document.emobg-sso-billing-admin.json
},
{
name = "emobg-sso-billing-audit",
inline_policy = data.aws_iam_policy_document.emobg-sso-billing-audit.json
},
{
... [ All MY POLICIES ARE LISTED HERE ]
}
]
}
I am creating some resources using the for_each method on Terraform version 0.14.15. The resource has an attirbute, input_parameters that takes a string in JSON format as its value. I am defining this value in a map variable utilizing separate objects. The value I am specifying as a string in JSON format, and I am getting an error upon execution that I need to declare a string. Any insight on fixing this error would be helpful. Below is how I have my resource and variable declared.
Resource
resource "aws_config_config_rule" "managed_rules" {
for_each = var.managed_rules
name = each.value.name
description = each.value.description
input_parameters = each.value.input_parameters
source {
owner = each.value.owner
source_identifier = each.value.source_identifier
}
depends_on = [aws_config_configuration_recorder.config_recorder]
}
Variable
variable "managed_rules" {
type = map(object({
name = string
description = string
owner = string
source_identifier = string
# Is there a variable for strings in JSON format?
input_parameters = string
}))
default = {
"1" = {
name = "alb-http-to-https-redirection-check"
description = "Checks whether HTTP to HTTPS redirection is configured on all HTTP listeners of Application Load Balancers. The rule is NON_COMPLIANT if one or more HTTP listeners of Application Load Balancer do not have HTTP to HTTPS redirection configured."
owner = "AWS"
source_identifier = "ALB_HTTP_TO_HTTPS_REDIRECTION_CHECK"
input_parameters = {
"MaximumExecutionFrequency" : "TwentyFour_Hours",
}
}
Error
This default value is not compatible with the variable's type constraint:
element "2": attribute "input_parameters": string required.
After updating code with jsonencode function and changing input_parameters to any, this is the error:
This default value is not compatible with the variable's type constraint:
collection elements cannot be unified.
You have a couple things going on here:
The resource requires input_parameters to be a JSON-encoded string
You have the variable type as a string
You're passing an object type into the variable that only accepts a string type
So (2) and (3) are conflicting. At some point, you have to convert your object into a JSON string. You can either do that before passing it in as an input variable, or change your input variable to accept objects and convert the object to JSON when providing it to the resource.
I'd choose the second option because it's more intuitive to pass the object into the module instead of a string. So, try this:
resource "aws_config_config_rule" "managed_rules" {
for_each = var.managed_rules
name = each.value.name
description = each.value.description
input_parameters = jsonencode(each.value.input_parameters)
source {
owner = each.value.owner
source_identifier = each.value.source_identifier
}
depends_on = [aws_config_configuration_recorder.config_recorder]
}
variable "managed_rules" {
type = map(object({
name = string
description = string
owner = string
source_identifier = string
# Is there a variable for strings in JSON format?
input_parameters = any
}))
default = {
"1" = {
name = "alb-http-to-https-redirection-check"
description = "Checks whether HTTP to HTTPS redirection is configured on all HTTP listeners of Application Load Balancers. The rule is NON_COMPLIANT if one or more HTTP listeners of Application Load Balancer do not have HTTP to HTTPS redirection configured."
owner = "AWS"
source_identifier = "ALB_HTTP_TO_HTTPS_REDIRECTION_CHECK"
input_parameters = {
"MaximumExecutionFrequency" : "TwentyFour_Hours",
}
}
Note that I've used jsonencode in the resource's input_parameters and I've changed the variable type for that field to any (so it will accept an object of any structure).
You can create your json string as follows:
variable "managed_rules" {
type = map(object({
name = string
description = string
owner = string
source_identifier = string
# Is there a variable for strings in JSON format?
input_parameters = string
}))
default = {
"1" = {
name = "alb-http-to-https-redirection-check"
description = "Checks whether HTTP to HTTPS redirection is configured on all HTTP listeners of Application Load Balancers. The rule is NON_COMPLIANT if one or more HTTP listeners of Application Load Balancer do not have HTTP to HTTPS redirection configured."
owner = "AWS"
source_identifier = "ALB_HTTP_TO_HTTPS_REDIRECTION_CHECK"
input_parameters = <<EOL
{
"MaximumExecutionFrequency" : "TwentyFour_Hours",
}
EOL
}
}
}
But then you have to use jsondecode if you want to parse this string. You can't use functions in variables, so it must be done later.
To add to Jordan's answer. I Had a similar concern when trying to add a json policy to a module.
I used the any object type in place of the string object type.
Here's how I fixed it:
Module main.tf
resource "aws_ecr_repository_policy" "main" {
repository = var.repository_name
policy = var.repository_policy
}
Module variables.tf
variable "repository_name" {
type = string
description = "Name of the repository."
}
variable "repository_policy" {
type = any
description = "The policy document. This is a JSON formatted string."
}
Resource Creation main.tf
# ECR Repository for container images
module "ecr_repository_1" {
source = "../../../../modules/aws/ecr-repository"
ecr_repository_name = var.ecr_repository_name.1
image_tag_mutability = var.image_tag_mutability
image_scan_on_push = var.image_scan_on_push
tag_environment = local.tag_environment
tag_terraform = local.tag_terraform.true
}
# ECR Repository policies
module "ecr_repository_policy_1" {
source = "../../../../modules/aws/ecr-repository-policy"
repository_name = var.ecr_repository_name.1
repository_policy = var.repository_policy.1
}
Resource creation variables.tf
variable "ecr_repository_name" {
type = map(string)
description = "Name of the repository."
default = {
"1" = "my-backend-api"
}
}
variable "image_tag_mutability" {
type = string
description = "The tag mutability setting for the repository. Must be one of: MUTABLE or IMMUTABLE. Defaults to MUTABLE."
default = "MUTABLE"
}
variable "image_scan_on_push" {
type = bool
description = "Indicates whether images are scanned after being pushed to the repository (true) or not scanned (false)."
default = true
}
variable "repository_policy" {
type = any
description = "The policy document. This is a JSON formatted string."
default = {
"1" = <<EOF
{
"Version": "2008-10-17",
"Statement": [
{
"Sid": "new policy",
"Effect": "Allow",
"Principal": "*",
"Action": [
"ecr:GetDownloadUrlForLayer",
"ecr:BatchGetImage",
"ecr:BatchCheckLayerAvailability",
"ecr:PutImage",
"ecr:InitiateLayerUpload",
"ecr:UploadLayerPart",
"ecr:CompleteLayerUpload",
"ecr:DescribeRepositories",
"ecr:GetRepositoryPolicy",
"ecr:ListImages",
"ecr:DeleteRepository",
"ecr:BatchDeleteImage",
"ecr:SetRepositoryPolicy",
"ecr:DeleteRepositoryPolicy"
]
}
]
}
EOF
}
}
input_parameters = {
"MaximumExecutionFrequency" : "TwentyFour_Hours",
}
This has to be a string instead of Object, Since you defined it as String
How do you capture a JSON object as a prettified string using a Jenkins declarative-syntax pipeline?
pipeline {
agent any
stages {
stage( "Set up" ) {
steps {
script {
hostname = "bld-machine"
reply_email = "jenkins#${hostname}.company.com"
actor_email = "user#company.com"
status_json = initStatusJson()
}
}
}
/** Try figure out the difference between "global" and "env." variables. */
stage( "Capture variables" ) {
steps {
script {
status_json.env["var"] = "${env.var}" as String
status_json.env["var2"] = "${var}" as String
}
}
}
}
post {
always {
script {
def pretty_json = writeJSON( returnText: true, json: status_json )
}
emailext( subject: "CI/CD | ${currentBuild.currentResult}",
from: "${reply_email}",
to: "${actor_email}",
mimeType: "text/plain",
body: "${pretty_json}" )
}
}
}
def initStatusJson() {
def json_obj = readJSON text: '{}'
json_obj.job = readJSON text: '{}'
json_obj.env = [:]
json_obj.job.name = "${JOB_BASE_NAME}" as String
json_obj.job.number = "${BUILD_ID}" as String
json_obj.job.server = "${JENKINS_URL}" as String
json_obj.job.visualization = "${JENKINS_URL}/blue/organizations/jenkins/${JOB_BASE_NAME}/detail/${JOB_BASE_NAME}/${BUILD_ID}/pipeline" as String
return json_obj
}
The def pretty_json =... statement in the above Jenkinsfile triggers the following error:
WARNING: Unknown parameter(s) found for class type WARNING: Unknown parameter(s) found for class type 'org.jenkinsci.plugins.pipeline.utility.steps.json.WriteJSONStep': returnText
[Pipeline] }
[Pipeline] // script
Error when executing always post condition:
java.lang.IllegalArgumentException: You have to provided a file for writeJSON.
at org.jenkinsci.plugins.pipeline.utility.steps.json.WriteJSONStepExecution.run(WriteJSONStepExecution.java:61)
at org.jenkinsci.plugins.pipeline.utility.steps.json.WriteJSONStepExecution.run(WriteJSONStepExecution.java:43)
at org.jenkinsci.plugins.workflow.steps.SynchronousNonBlockingStepExecution.lambda$start$0(SynchronousNonBlockingStepExecution.java:47)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
at java.util.concurrent.FutureTask.run(FutureTask.java:266)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
What I have tried:
The def pretty_json = writeJSON( returnText: true, json: status_json ) statement is inspired by these resources:
Jenkinsfile pipeline construct JSON object and write to file
https://www.jenkins.io/doc/pipeline/steps/pipeline-utility-steps/#writejson-write-json-to-a-file-in-the-workspace
I also tried def pretty_json = writeJSON returnText: true, json: status_json which resulted in an identical error.
status_json.toString() returns a valid, but non-prettified JSON string.
I tried def pretty_json = JsonOutput.toJson(status_json) based on Create JSON strings from Groovy variables in Jenkins Pipeline, and it generates this error:
Error when executing always post condition:
groovy.lang.MissingPropertyException: No such property: JsonOutput for class: groovy.lang.Binding
Tried def pretty_json = groovy.json.JsonOutput.prettyPrint(status_json) based on https://gist.github.com/osima/1161966, and it generated this error:
Error when executing always post condition:
groovy.lang.MissingMethodException: No signature of method: java.lang.Class.prettyPrint() is applicable for argument types: (net.sf.json.JSONObject)
Update: Attempted #daggett's solution as follows:
post {
always {
script {
def pretty_json = status_json.toString(2)
}
emailext( subject: "CI/CD | ${currentBuild.currentResult}",
from: "${reply_email}",
to: "${actor_email}",
mimeType: "text/plain",
body: "${pretty_json}" )
}
}
...and also tried some variations like pretty_json = ... (instead of def pretty_json = ...) and also moving the pretty_json assignment outside the script{...} scope...but none worked.
Inside the script{...} context, the .toString(2) generated this error:
Scripts not permitted to use method net.sf.json.JSON toString int.
Outside the script{...} context, it generated what I interpret to be a "syntax error":
WorkflowScript: 79: Expected a step # line 79, column 7.
pretty_json = status_json.toString(2)
According to last error message
groovy.lang.MissingMethodException:
No signature of method: java.lang.Class.prettyPrint()
is applicable for argument types: (net.sf.json.JSONObject)
You have net.sf.json.JSONObject in status_json variable.
that's really strange - seems you are getting status_json not in a standard way for jenkins
however according to documentation of this class
http://json-lib.sourceforge.net/apidocs/jdk15/net/sf/json/JSONObject.html#toString(int)
just do following to make pretty json:
def pretty_json = status_json.toString(2)
If you have Scripts not permitted to use method XYZ exception:
for security reasons a lot of non-standard methods are disabled in jenkins.
refer this answer to resolve this kind of issue: https://stackoverflow.com/a/39412951/1276664
and finally - almost every case from your question should work:
writeJSON( returnText: true, json: status_json ) :
update pipeline-utility-steps jenkins plugin to the latest version to support returnText parameter
the same as above
...
JsonOutput.toJson(status_json) : JsonOutput class located in groovy.json package. you could import this package at t
he beginning of the script import groovy.json or call it like this: groovy.json.JsonOutput.toJson(status_json). note that this method returns non-formatted json.
groovy.json.JsonOutput.prettyPrint(status_json) : check the documentation for JsonOutput.prettyPrint - it could be called for string and not for object. so this could work: groovy.json.JsonOutput.prettyPrint(status_json.toString()) but only in case when status_json.toString() returns a valid json and JsonOutput.prettyPrint allowed to be called in jenkins admin.
I just did a test and it gave results :
def pretty_json = writeJSON( returnText: true, json: status_json , pretty: 4)
Note : Ensure you have the plugin Pipeline Utility Steps installed. Or reinstall it again.
Below is the script example:
#!groovy
import hudson.model.Result
import groovy.json.*
pipeline
{
agent any
stages
{
stage ('Set up')
{
steps
{
script
{
hostname = "bld-machine"
reply_email = "jenkins#${hostname}.company.com"
actor_email = "user#company.com"
status_json = initStatusJson()
println (status_json)
}
}
}
stage ('Capture variables')
{
steps
{
script
{
// Added just for test
status_json.env["var"] = "Alt" as String
status_json.env["var2"] = "su" as String
println (status_json)
}
}
}
}
post {
always {
script {
def pretty_json = writeJSON( returnText: true, json: status_json , pretty: 4)
println (pretty_json)
emailext( subject: "CI/CD | ${currentBuild.currentResult}",
from: "${reply_email}",
to: "${actor_email}",
mimeType: "text/plain",
body: "${pretty_json}" )
}
}
}
}
def initStatusJson() {
def json_obj = readJSON text: '{}'
json_obj.job = readJSON text: '{}'
json_obj.env = [:]
json_obj.job.name = "${JOB_BASE_NAME}" as String
json_obj.job.number = "${BUILD_ID}" as String
json_obj.job.server = "${JENKINS_URL}" as String
json_obj.job.visualization = "${JENKINS_URL}/blue/organizations/jenkins/${JOB_BASE_NAME}/detail/${JOB_BASE_NAME}/${BUILD_ID}/pipeline" as String
return json_obj
}
Output log :
I'm new to LUA and tried learning coding this language with Garrys Mod.
I want to get the messages from the Garrys Mod chat and send them into a Discord channel with a webhook.
It works, but I tried expanding this project with embeded messages. I need JSON for this and used json.lua as a library.
But as soon as I send a message I retrieve the following error message:
attempt to index global 'json' (a nil value)
The code that causes the error is the following:
json.encode({ {
["embeds"] = {
["description"] = text,
["author"] = {
["name"] = ply:Nick()
},
},
} }),
The complete code:
AddCSLuaFile()
json = require("json")
webhookURL = "https://discordapp.com/api/webhooks/XXX"
local DiscordWebhook = DiscordWebhook or {}
hook.Add( "PlayerSay", "SendMsg", function( ply, text )
t_post = {
content = json.encode({ {
["embeds"] = {
["description"] = text,
["author"] = {
["name"] = ply:Nick()
},
},
} }),
username = "Log",
}
http.Post(webhookURL, t_post)
end )
I hope somebody can help me
Garry's Mod does provide two functions to work with json.
They are:
util.TableToJSON( table table, boolean prettyPrint=false )
and
util.JSONToTable( string json )
There is no need to import json and if I recall correctly it isn't even possible.
For what you want to do you need to build your arguments as a table like this:
local arguments = {
["key"] = "Some value",
["42"] = "Not always the answer",
["deeper"] = {
["my_age"] = 22,
["my_name"] = getMyName()
},
["even more"] = from_some_variable
and then call
local args_as_json = util.TableToJSON(arguments)
Now you can pass args_as_json to your
http.Post( string url, table parameters, function onSuccess=nil, function onFailure=nil, table headers={} )
I'd like to create a JSON object in Swifty that has the form:
{
"store": {
"id": {
"test": "test"
},
"type": "retail",
"name": "store1"
}
}
Is there a way to combine types in a Dictionary to use with Swifty (String and JSON)? Quotes works, but when I try to assign a variable, it complains: Cannot assign value of type 'String' to type 'JSON?':
func jsonTest()->String {
var storeJson = [String: JSON]()
var someJson = JSON(["test":"test"])
storeJson["id"] = someJson
storeJson["type"] = "retail" // <-- works fine
var name = "store1"
storeJson["name"] = name // <-- Doesn't work
var store = JSON(storeJson)
return store.rawString()!
}
The reason
storeJson["type"] = "retail"
works differently than
storeJson["name"] = name
is because the first one follows a different path in the code. Specifically, it uses the init(stringLiteral value: StringLiteralType) method in the following extension (source).
extension JSON: Swift.StringLiteralConvertible {
public init(stringLiteral value: StringLiteralType) {
self.init(value)
}
public init(extendedGraphemeClusterLiteral value: StringLiteralType) {
self.init(value)
}
public init(unicodeScalarLiteral value: StringLiteralType) {
self.init(value)
}
}
I'll explain further after we talk about how to fix your specific problem.
Possible solution #1:
storeJson["name"]?.string = name
Output:
{
"id" : {
"test" : "test"
},
"type" : "retail"
}
The reason
storeJson["name"]?.string = name
doesn't work as we might think is because of the optional chaining. Right now, if we ran this through the debugger, we wouldn't see anything meaningful. In fact, we would see nothing. This is a bit concerning and likely means storeJson["name"] is nil, so the statement is not executing any further. Let's verify our hypothesis by making it blow up. We'll change the line to:
storeJson["name"]!.string = name
In this case, with your current code, you'll likely get
fatal error: unexpectedly found nil while unwrapping an Optional value
as you should because storeJson["name"] is in fact nil. Therefore, this solution doesn't work.
Possible solution #2:
As you correctly noted in your answer, if you add a storeJson["name"] = JSON(name), you'll get the desired behavior:
func jsonTest()->String {
var storeJson = [String: JSON]()
var someJson = JSON(["test":"test"])
storeJson["id"] = someJson
storeJson["type"] = "retail" // <-- works fine
var name = "store1"
storeJson["name"] = JSON(name) // <-- works!
var store = JSON(storeJson)
return store.rawString()!
}
Output:
{
"id" : {
"test" : "test"
},
"name" : "store1",
"type" : "retail"
}
Great! Therefore, this solution works! Now, later in your code you can alter it however you want using .string and the like.
Explanation
Back to why the string literal works. You'll notice in the init, it has
self.init(value)
which passes through the objects init, which then goes through the case statement
...
case let string as String:
_type = .String
self.rawString = string
...
When you call storeJson["name"] = JSON(name), you're skipping the StringLiteralType init and simply going into the switch.
Therefore, you could interchange
storeJson["type"] = "retail"
with
storeJson["type"] = JSON("retail")
It turns out it works to change:
storeJson["name"] = name
to
storeJson["name"] = JSON(name)