Related
Our project has a few GRPC servers defined as go_binary targets. We develop client SDKs for Java and Python applications and we would like to use java_test and py_test. Is any way to start a specific go_binary target before java_test or py_test?
You can create a test harness that starts the gRPC server before running the tests. For example, you could add the binary to the data attribute of the test, and then started it beforehand:
go_binary(
name = "my_grpc_server",
[...]
)
py_test(
name = "my_test",
[...]
data = [":my_grpc_server"],
)
and then inside the test file:
class ClientTestCase(unittest.TestCase):
def setUp(self):
r = runfiles.Create()
self.server = subprocess.Popen([r.Rlocation("path/to/my_grpc_server")])
def tearDown(self):
self.server.terminate()
self.server.wait()
This example is very simple, you'll probably run into issues regarding the availability of the port the server listens on, or waiting for the server to start up. You could add flags to your gRPC server to allow communication over a domain socket, or make it listen on an unused port and have the test parse the port number from the server's log output.
For details on finding the server with runfiles: https://github.com/bazelbuild/bazel/blob/a7a0d48fbeb059ee60e77580e5d05baeefdd5699/tools/python/runfiles/runfiles.py#L16-L58
If you find yourself copy-pasting this pattern a lot, or having to implement it in multiple languages, you could try using an sh_test() rule to wrap the underlying py_test or java_test, and to start the server, then start the test with an environment variable telling it how to reach the server (eg MY_GRPC_SERVER_ADDRESS=localhost:${test_port}.
In Bazel, how do I fetch a remote file as a build rule not as a WORKSPACE rule?
I want to use a build rule because WORKSPACE rules are not loaded for transitively.
e.g. this fails
load("#bazel_tools//tools/build_defs/repo:http.bzl", "http_file")
http_file(
name = "foo",
urls = [ "https://example.com" ],
sha256 = "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
executable = True,
)
Error in repository_rule: 'repository rule http_file' can only be called during workspace loading
If you really want to do that, you have to implement your own rule, a naïve trivial example relying on curl to fetch could be:
def _impl(ctx):
args = ctx.actions.args()
args.add("-o", ctx.outputs.out)
args.add(ctx.attr.url)
ctx.actions.run(
outputs = [ctx.outputs.out],
executable = "curl",
arguments = [args],
)
get_stuff = rule(
_impl,
attrs = {
"url": attr.string(
mandatory = True,
),
},
outputs = {"out": "%{name}.out"},
)
But (and esp. in such a trivial) for, it comes with problems. Apart from, do you want to step out of sandbox during the build? And do you want to talk to someone across the network during the build (out of the sandbox)? Bypassing repository_cache, and possibly getting remote_cache involved (networked caching of networked fetching). Specifically in this example, if content of the file pointed to by url changes... build has no idea and only fetches it when it either hasn't done so or the url itself has changed. I.e. the implementation would need to be more robust (mimic that of http_file for instance).
But it actually sounds like you're trying to address a different problem (transitive external dependencies, for which there could be another solution). One trick used for that is to define a macro (in your first level dependency to load define the next hop) and after declaring that first step as an external dependency in your parent project, load the that macro and use it from parent project WORKSPACE. This too has a price though, namely the first level dependency has to always be present (fetched or already cached), even if build target asked for does not actually need it (as that load and macro call will always pull it in).
I have the following three files as below:
main.tf, variables.tf and dev.auto.tfvars
Snippet from main.tf
module "sql_vms" {
source = "git::git#github.com:xxxxxxxxxxxx/terraform-modules//azure/"
rg_name = var.resource_group_name
location = module.resource_group.external_rg_location
vnet_name = var.virtual_network_name
subnet_name = var.sql_subnet_name
app_nsg = var.application_nsg
vm_count = var.count_vm
base_hostname = var.sql_host_basename
sto_acc_suffix = var.storage_account_suffix
vm_size = var.virtual_machine_size
vm_publisher = var.virtual_machine_image_publisher
vm_offer = var.virtual_machine_image_offer
vm_sku = var.virtual_machine_image_sku
vm_img_version = var.virtual_machine_image_version
username = var.username
password = var.password
}
Snippet from variables.tf
variable "app_subnet_name" {
type = string
}
variable "sql_subnet_name" {
type = string
}
Snippet from dev.auto.tfvars
app_subnet_name = "subnet_1"
sql_subnet_name = "subnet_2"
application_nsg = "test_nsg"
However, I'm getting error like below
Error: Unsupported argument
on main.tf line 7, in module "sql_vms":
7: subnet_name = var.sql_subnet_name
An argument named "subnet_name" is not expected here.
Error: Unsupported argument
on main.tf line 8, in module "sql_vms":
8: app_nsg = var.application_nsg
An argument named "app_nsg" is not expected here.
My modules directory structure looks like below
$ ls -R terraform-modules/
terraform-modules/:
aws azure gcp
terraform-modules/aws:
alb ec2-instance-rhel
terraform-modules/aws/alb:
terraform-modules/aws/ec2-instance-rhel:
main.tf
terraform-modules/azure:
compute resourcegroup sqlserver
terraform-modules/azure/compute:
main.tf README.md variable.tf
terraform-modules/azure/resourcegroup:
data.tf outputs.tf variables.tf
terraform-modules/azure/sqlserver:
main.tf README.md variables.tf
terraform-modules/gcp:
compute
terraform-modules/gcp/compute:
main.tf
Any idea what is going wrong here?
If you are starting out with Terraform, you will get that error message ("An argument named "example" is not expected here") if your module arguments refer to the resource properties and not to variable names, see below for an example:
Example of a Terraform module "example_mod.tf" you want to call from your module:
variable "sg_name" { } # Usually in a separate file
variable "sg_desc" { } # called variables.tf
resource "example_resource" "example_name" {
name = var.sg_name
description = var.sg_desc
...
}
CORRECT WAY:
module "my_module" {
source = "./modules/example_mod.tf"
sg_name = "whatever" # NOTE the left hand side "sg_name" is the variable name
sg_desc = "whatever"
...
}
INCORRECT WAY: (Gives the error "An argument named "name" is not expected here" )
module "my_module" {
source = "./modules/example_mod.tf"
name = "whatever" # WRONG because the left hand side "name" is a resource property
description = "whatever" # WRONG for the same reason
...
}
I think the issue is that you do not refer to the exact module with the source. I see you have three modules in the source:
source = "git::git#github.com:xxxxxxxxxxxx/terraform-modules//azure/"
They are compute, resourcegroup and sqlserver. But you want to load them in one module. So it cannot find the related variables for the modules. I also don't think it's the right way to load all the modules like that. I would recommend you load the modules one by one like below:
module "compute" {
source = "git::git#github.com:xxxxxxxxxxxx/terraform-modules//azure/compute"
...
}
module "resourcegroup" {
source = "git::git#github.com:xxxxxxxxxxxx/terraform-modules//azure/resourcegroup"
...
}
module "sqlserver" {
source = "git::git#github.com:xxxxxxxxxxxx/terraform-modules//azure/sqlserver"
...
}
Without knowing the details about the module it is usually hard to say what's the reason for an error, but in this particular case it seems that there isn't a requirement in the module you're importing to use those two arguments (subnet_name and app_nsg), or rather that you are using a version of the module that doesn't require them to be present. What helps with that type of error is to check if there is a version of the module that does have such a requirement. The syntax for using a particular module version from Github is explained in Terraform Module Sources documentation, Selecting a Revision section:
module "vpc" {
source = "git::https://example.com/vpc.git?ref=v1.2.0"
}
You are probably using SSH to fetch the module, so the recommended way to do that is:
When using Git over SSH, we recommend using the ssh://-prefixed URL form for consistency with all of the other URL-like git address forms.
In your example, this translates to:
module "sql_vms" {
source = "git::ssh://git#github.com/org/terraform-modules-repo.git//azure/module-name?ref=v1.2.0"
where org is your organisation's (or your private) Github account, terraform-modules-repo is the repo where modules reside, module-name is the module you are using and ref=v1.2.0 represents the module revision number.
The error An argument named "example" is not expected here. means that the module doesn't expect to see an input argument with that name. Think about Terraform modules as functions in a programming language: in order to have a function provide a result, you pass the function a set of required arguments. If you provide more (or less) input arguments than required by that function call, you will get an error. (There are special cases but it is out of the scope of this question.)
Another similarity between modules and functions is that Terraform modules can also provide output values, besides creating resources that are specified. That can be handy in cases where output can be used as input in other modules or resources. The line module.resource_group.external_rg_location is doing exactly that: getting the output value from another module and using it to assign a value to an argument location.
I had a similar issue when working with AWS Eventbridge and Terraform.
When I run terraform plan I get the error below:
Error: Unsupported argument
│
│ on ../../modules/aws/eventbridge/main.tf line 37, in resource "aws_cloudwatch_event_target" "ecs_cloudwatch_event_target":
│ 37: maximum_age_in_seconds = var.maximum_age_in_seconds
│
│ An argument named "maximum_age_in_seconds" is not expected here.
Here's how I solved it:
The issue was that I was not using the correct attribute for the AWS Eventbridge resource block.
The attribute should have been maximum_event_age_in_seconds and not maximum_age_in_seconds.
Another issue that could this is not defining a variable in your terraform script that is already defined in a module.
That's all
It could be happening due to plenty of reasons.
I'd suggest some verification:
Check if you are using the correct source URL, path or revision branch/tag.
I'm not sure about your implementation, but you probably want to double check the revision you are referencing contains theses variable declarations.
GitHub Modules addressing allows ref argument.
Refer to the GitHub Module Addressing for Terraform Documentation and how to specify a revision.
Check if all necessary variables are declared on every module, including the root module.
Did you declare those variables both in a variables.tf file on your root directory and on the module context/path?
I know that's exhausting and repetitive, but every module should be designed as an "independent project". Each module **MUST have its own declared variables.tf**, which work as inputs for that module, and it is also desirable that it has its own mapped outputs.tf, provider.tf, backend.tf, etc., though these last ones are not required.
FYI: Doing so you guarantee scalability, reusability, as well as reliability to work with different tfstate files and even different repositories for each module in order to guarantee atomicity and minimum permissions, hence preventing your infrastructure from being destroyed by undesired code changes.
I highly recommend this read to understand the importance of independent modularization design.
Furthermore, tools like Terragrunt, Terratest can make this job less painful by keeping your code DRY ( Don't Repeat Yourself ).
Check if the **type constraints of the related variables match.**
If that's not your case, try looking if the type constraints match between all declarations of the variables used both as arguments ( on your root variables.tf ) and inputs ( on your module level variables.tf ).
I'll share my pain as well.
Writing block configuration like this
vpc_config = {
subnet_ids = [aws_subnet.example1.id, aws_subnet.example2.id]
}
Instead of (Notice the = Equal Sign):
vpc_config {
subnet_ids = [aws_subnet.example1.id, aws_subnet.example2.id]
}
Will give an error of An argument named "vpc_config" is not expected here and will waste you a few good hours.
I'm working on implementing a feature like Strict Java Deps for rules_scala.
I'd really like to have the ability to configure in runtime if this uses warn or error.
I seem to recall skylark rules can't create and access command-line flags but I don't recall if they can access existing ones?
Main difference is that existing ones are already parsed so maybe they are also passed in some ctx.
The flag you want (strict_java_deps) isn't available through Skylark at the moment. There's no reason we can't add it, though, filed #3295 to track.
For other flags, the context can access the configuration fragments, which can access some of the parsed command line flags. I think what you'd want is ctx.fragments, then use the fragments to get the java fragments, and then get the default_javac_flags from that:
# rules.bzl
def _impl(ctx):
print("flags: %s" % ctx.fragments.java.default_javac_flags)
...
frag = rule(
implementation = _impl,
fragments = ["java"], # Declare that this rule uses java fragments
)
Then:
$ bazel build --javacopt="-g:source,lines" :x
WARNING: /home/kchodorow/test/a/tester.bzl:2:3: flags: ["-g:source,lines"].
I have been able to successfully configure resource environment entries with the Jython script below. I call the Jython script with the ws_admin program in my local app servers bin directory.
I work on a team where ant is the preferred technology in our build process.
I've looked around the web for documentation on configuring WebSphere with ant and so far it looks like to me that one is mainly able to call programs like ws_admin from ant.
Is it possible to configure resource environment entries using ant directly instead of using a Jython or Jacl script? If not, how can I go about setting up an ant task to reduce the amount of Jython that is needed to set up resource environment entries?
Here's my current Jython script that sets up resource environment entries. Ultimately looking for ways to reduce our dependence on Jython...
# Set up Variables used within this script
objServerAttrs = AdminControl.completeObjectName('WebSphere:type=Server,*')
node = AdminControl.getAttribute(objServerAttrs, 'nodeName')
server = AdminControl.getAttribute(objServerAttrs, 'name')
provider = "Test_ConfigurationProvider"
providerFactory = "com.DG_ConfigurationFactory"
providerClass = "com.DG_Configuration"
# Function for creating resource custom properties
def createResourceCustomProperty(envEntry, propName, propValue):
propSet = AdminConfig.showAttribute(envEntry, 'propertySet')
if propSet == None:
propSet = AdminConfig.create('J2EEResourcePropertySet',envEntry,[])
name = ['name', propName]
value = ['value', propValue]
propAttrs = [name, value]
AdminConfig.create('J2EEResourceProperty', propSet, propAttrs)
return
# Create the resource environment provider
AdminResources.createResourceEnvProvider(node, server, provider)
AdminResources.createResourceEnvProviderRef(node,server,provider, providerFactory, providerClass)
# Create the resource environment entries
## Context Configuration
envEntry = AdminResources.createResourceEnvEntries(node,server,provider, "Context Configuration", "test-config/context")
createResourceCustomProperty(envEntry, "deployment.environment", "IDE")
createResourceCustomProperty(envEntry, "server.context", "com.context.DG_WebSphereServerContext")
createResourceCustomProperty(envEntry, "user.context", "com.context.DG_WebSphereUserContext")
createResourceCustomProperty(envEntry, "log.directory", "C:/Development/WebSphere/Logs")
createResourceCustomProperty (envEntry, "file.directory", "C:/Development/WebSphere/AppFiles")
## Mail Configuration
envEntry = AdminResources.createResourceEnvEntries(node,server,provider, "Mail Configuration", "test-config/mail")
createResourceCustomProperty(envEntry, "enabled", "false")
createResourceCustomProperty(envEntry, "mailSessionJndiName", "mail/MailSession")
## User Repository Configuration
envEntry = AdminResources.createResourceEnvEntries(node, server, provider, "User Repository Configuration", "test-config/userRepository")
createResourceCustomProperty(envEntry, "ldap.provider.url", "ldap://test.com:389/cn=users,dc=com")
createResourceCustomProperty (envEntry, "ldap.security.principal", "cn=was_user,cn=users,dc=com")
# Save changes to the configuration
AdminConfig.save()
Starting with WAS 7, in addition to admin console and wsadmin, a third way to configure the server was introduced, namely properties file based configuration. This new administrative model supposedly "eliminates the need to write complex wsadmin scripts" as explained at related Education Assistant presentation.
What you do, basically, is configure a single environment, export the parts of configuration that are of interest to a portable properties file, and later use this file as an input to a single line of wsadmin script, which applies the configuration in the properties file to another target server. So you get rid of many lines of Jython and work with a much simpler artefact, which is a property file with a simple and familiar syntax.
In addition to above links there is a nice article about this feature at Developerworks.