Hey there, getting started and I tried using the container-aws-typescript template. When I run

pulumi up

I get an error from aws:ec2/getVpc "no matching EC2 VPC found". Should I expect the template to yield a deployable starting point?

According to the docs at https://www.pulumi.com/templates/container-service/aws/, you should:

When it’s done, you’ll have a complete Pulumi project that’s ready to deploy and configured with the most common settings.

There isn't a call to getVpc() in the template / generated code. That's getting called, I assume, from each resource.

awsx (which is used in this template) used the default vpc, it looks like you don’t have any default vpc?

Should I expect the template to yield a deployable starting point?

Definitely should, yeah, and sounds like this may be because of a missing default VPC. I haven't tried this myself, but you should be able to make a few tweaks to the template to create a new VPC and then use it by following the example here: https://www.pulumi.com/docs/clouds/aws/guides/ecs/#creating-an-ecs-cluster-in-a-vpc

hi from me too! 💜

Update: Confirmed with a colleague that the default VPC had previously been deleted from the region I was running in. I proceeded using vanilla the aws provider in the meantime.

The DefaultVpc resource is quite good for this. It will create a new default VPC or adopt the existing default VPC, no special code required. It's in the aws.ec2 package.

I think I want to start with creating my own as that's what I'll need later, but good to know there's such an option to get or create a default!

Use the vpc property of your vpc object; that's the aws.ec2 Vpc.

Hmm, I'm not sure if I follow. However, the following "compiles"

const alb = new <http://awsx.lb|awsx.lb>.ApplicationLoadBalancer("web-traffic", vpc, {});

Passing vpc as an arg rather than in the options. That is not what the doc shows. It seems this isn't the only doc discrepancy as later the guide proposes:

const listener = alb.createListener("web-listener", { port: 80 });

But that method doesn't exist.

Argh, looks like this doc is outdated, yeah. I'll see if I can dig up a working example that does what you need here. FWIW, this doc corresponds with the beta version of

awsx

, which recently went 1.0 (and unfortunately changed a few APIs). Those APIs are still there, but they're available under the

classic

namespace -- so when you run into this, you can generally switch your references from

awsx.module.ResourceName

to

awsx.classic.module.ResourceName

to use them, if that helps. Meantime though, I'll see if I can find something that does all this with the 1.0 SDK. (And file to get that doc updated!)

Thanks! Looking at it more I don't understand why passing vpc as second argument (as above) works. That second arg should be an ApplicationLoadBalancerArgs interface 🙃 I think I found a work around for creating the lb in the custom vpc by specifying subnets to use:

const alb = new <http://awsx.lb|awsx.lb>.ApplicationLoadBalancer(rname("web-traffic"), {

subnetIds: vpc.publicSubnetIds,

});

Yep, that looks right. I have something working that I can share -- at least it's working for me 🙂 https://gist.github.com/cnunciato/812531c537ccf9a0e94674d11cd62ef7 I'm on an M2, hence that

DOCKER

env var (without that, the images I build locally don't run when they make it to AWS), and you'll probably want to adjust the subnet group (as it allows everything right now), but hopefully this'll help ya.

Thanks so much! Using your example, I tweaked what I had and finally got to a reachable nginx stub. I need to stop for the day, but thanks again for the help.