bitter-dentist-28132
12/18/2020, 8:11 PMerror: Error: Command failed: helm template /tmp/tmp-31838777XOYQC967QQm/deprecation --name-template nginxingresscontroller --values /tmp/tmp-31838777XOYQC967QQm/deprecation/values.yaml --values /tmp/tmp-31838771tvGEyQxKeRB.yaml --namespace default
Error: open /tmp/tmp-31838777XOYQC967QQm/deprecation/values.yaml: no such file or directory
anyone have any ideas?dry-engine-17210
12/18/2020, 9:03 PMjolly-vr-53742
12/18/2020, 10:41 PMjolly-vr-53742
12/18/2020, 11:05 PMproud-pizza-80589
12/19/2020, 5:17 PMbillions-oxygen-70106
12/21/2020, 1:07 AMpulumi up
- it works fine and there are no changes
3. I import a second resource (status shows imported)
4. When running pulumi up
, I get a duplicate URN error for the stack:
error: Duplicate resource URN 'urn:pulumi:org::some-project::pulumi:pulumi:Stack::some-project-org'; try giving it a unique name
Does anyone have an idea what I may be doing wrong? I've been fiddling with this for a while.worried-queen-62794
12/21/2020, 7:24 AMrefresh
to show me the differences? details
doesn't show anything. If I run pulumi up
then all is well but then an immediate pulumi refresh
shows changes.swift-hamburger-98290
12/21/2020, 11:29 AMapp-something
, which gets created with the name app-something{suffix}
, instead of app-something-{suffix}
(notice the missing -
). Dotnet API, on Azure. Thank you.magnificent-restaurant-51456
12/21/2020, 3:29 PMjolly-vr-53742
12/21/2020, 3:49 PMbillions-yak-67755
12/21/2020, 5:16 PMerror: aws:eks/cluster:Cluster resource 'training-eks-cluster' has a problem: Computed attribute cannot be set
Adding debug / verbose logs has not provided anything that seemed useful, and none of the many variants to parameters (providing, omitting, computed, hardcoded, etc.) have provided any progress, either. Any ideas on next steps for digging in and figuring out if this is user-error or a bug?hallowed-beard-11153
12/21/2020, 8:10 PMworried-queen-62794
12/22/2020, 4:48 AMpulumi up
in one of them it says the name from Pulumi.yaml doesn't match the stack's project name. If I rename it to match then the other one has the problem. Where is this information stored? Something must be referring back to this directory which seems strange.shy-toddler-33930
12/22/2020, 10:10 AMprehistoric-kite-30979
12/22/2020, 1:54 PMfuture-oyster-5387
12/22/2020, 9:21 PMroutes.push({
method: 'ANY',
path: '/',
localPath: path.join(__dirname, '../', 'frontend', 'build'),
})
but I don't see any options for it. I could set the content of /index.html to the 404 error page from s3, but I also don't obviously see how to do this.faint-dress-91709
12/22/2020, 10:45 PMconst tokenSecret = pulumi.output(args.tokenSecretArn).apply(secretArn =>
aws.secretsmanager.getSecretVersion({
secretId: secretArn,
})
)
billions-yak-67755
12/23/2020, 12:01 AMOutput.all()
(in Python) creates a list. Is it possible to have some variant of Output.all()
that creates a dict so keys / objects are easier to reference by name? If not, is this something it would be worth writing up as a feature request?
I.e.
# instead of
Output.all(output1, output2).apply(lambda x: x[0])
# something like
Output.all(name=output1, name2=output2).apply(lambda x: x['name'])
billions-yak-67755
12/23/2020, 1:59 AMOutput
objects, but most of that seemed to be within JavaScript. (i.e. if Output object attribute defined, then proceed, otherwise do not). Is this something do-able within python today? My invocations of Output.apply()
have thus far not been successful.
I.e. something like:
if (myoutput.apply(lambda x: x['arn']) !== ''):
# do something
https://github.com/pulumi/pulumi/issues/3942#issuecomment-588693999
Alternatively, I would like to schedule a "callback" to run within the Pulumi workflow between resources where I can call the Kubernetes API, etc.
Context - I need to deploy a configmap (for node registration) to an EKS cluster before I define the node group. I have been poking around pulumi_kubernetes
, but the auth to a newly created cluster seems that it will be a hangup (esp. with AWS and their iam-authenticator
business), and I already have a solution for this in native python. I just need to figure out how to get it into the Pulumi execution model. Should I do something like this inside of a big myoutput.apply()
(that feels wrong)?full-dress-10026
12/23/2020, 2:01 AMkubernetes:apps/v1:Deployment (my-deployment):
error: Preview failed: unexpected unknown property value
That's it. That's the extent of the message 🙂 I took the tried and true "comment things out until it works" approach, working it all the way back to a DeploymentArgs.spec.template.spec.containers[0].env[0].value (pod env value) being a pulumi.Input. This error was provoked by upgrading to pulumi 2.16.1 from 2.15.4. Not really sure what the fix is for this one.billions-oxygen-70106
12/23/2020, 6:45 AMTF_LOG=TRACE pulumi up -v9 --logtostderr
but there is not enough detail for me to debug what's going on under the hood.dry-engine-17210
12/23/2020, 6:16 PMbillions-yak-67755
12/23/2020, 9:12 PMkey:value
of <http://kubernetes.io/cluster/cluster-name:shared|kubernetes.io/cluster/cluster-name:shared>
.
However, this creates a chicken/egg problem, because the subnets must exist before the EKS cluster, but any subsequent runs will remove this tag and break the cluster.
Is there a way to apply tags "after the fact" to subnets? My workaround at present is to try to name the EKS cluster in a deterministic way so that I can set the tags on the subnet at create time. What I imagine is a aws.ec2.subnettag
provider or something of that sort... or if the aws.eks.cluster
provider somehow notified the subnet
provider that these tags should stayenough-butcher-66045
12/24/2020, 1:39 AM@Microsoft.KeyVault(SecretUri={slackBotTokenSecretUri})
format
My question is: I could obviously pass in that slack token via a parameter to the pulumi client, but where is it safely stored beforehand? I'm in a bit of a chicken/egg problem that I'm not quite sure how to solve.
Any help is appreciated, thank you!enough-butcher-66045
12/24/2020, 2:21 AMazure:appservice:AppService (WisrBotAppService):
error: transport is closing
pulumi:pulumi:Stack (WisrBot-dev):
panic: interface conversion: interface {} is nil, not map[string]interface {}
I have rolled back all my changes to a known state but I still get itrough-oxygen-8318
12/24/2020, 4:58 PMwet-sunset-4939
12/25/2020, 6:12 AMfuture-australia-56541
12/28/2020, 11:56 AMclever-plumber-29709
12/29/2020, 1:08 AMFrom that point onwards, Pulumi understands the inter-stack dependency for scenarios like cascading updates.
So in in project1 I export some resourceA-id.
then I import it in project2. and use it to deploy resourceB.
If i make changes on project1-resourceA, it may end-up changing its id.
So does pulumi automatically picks that for project B? or puts some output to warn about it?
do i manually have to run an pulumi up on project B for it to not break?
will i need to keep this changes in mind to avoid possible errors?? or is there a way to be more aware of this?acoustic-leather-88378
12/29/2020, 8:32 AMkms_key_id
etc. when logging in, as you would in the equivalent Terraform S3 backend below?
terraform {
backend "s3" {
bucket = "pulumi-state"
key = "test-state"
region = "eu-west-1"
encrypt = "true"
kms_key_id = "arn:aws:kms:eu-west-1:325425235:key/2345645645-210a-4ab6-a9ad-534536456"
acl = "private"
}
}