Error: Unable to get BasePathMappings for domain.com

promediacorp commented 5 years ago

Setting the string to empty as indicated in the examples, causes the the error.

Error: Unable to get BasePathMappings for domain.com

Also, no record is created in Route53

  customDomain:
    domainName: dev.domain.com
    basePath: ''
    certificateName: '*.domain.com'
    stage: ${self:provider.stage}
    createRoute53Record: true

Missing something?

felimartina commented 5 years ago

Having the same problem! Let me know if you were able to sort it out.

jguttman94 commented 5 years ago

The issue is in these lines.

It seems like the intention here is to create if non-existant, and update if existant...but instead, an error is being thrown in the helper method.

captainsidd commented 5 years ago

Hi there! Unfortunately, I'm not able to reproduce this error. Could you give me more information about how to reproduce it?

As @jackrg94 stated, we make the call to getBasePathMappings on the domain name in order to check if the basepath already exists. If the call to make the basepath fails (i.e. returns an error), we throw an error. Reasons why the getBasePathMappings call would fail are usually either related to the domain name not existing or incorrect permissions. Are either of those situations applicable in your cases?

We've heard from other users about how error messages aren't as descriptive as they'd like - we'll roll out a PR in the near future logging the error messages received from making AWS calls if SLS_DEBUG is set. We'll also be sure to update the README with the permissions required.

ryderjgillen commented 5 years ago

you can also get this error if you do not call sls create_domain

captainsidd commented 5 years ago

Hi everyone! We rolled out better debug messages in 3.2.1 that should help identify why you're experiencing this issue (incorrect permissions or not finding the domain as a result of not running sls create_domain as @rgillen stated). Let us know what's going on!

promediacorp commented 5 years ago

@captainsidd 😕 , still end up with Error: Unable to get BasePathMappings; simply trying to point route 53 A record alias to the cloudfront distribution.createRoute53Record: true doesnt seem to change anything in the dns. sls create_domain by itself yields Error: Failed to create custom domain

captainsidd commented 5 years ago

@promediacorp Hmmm, what happens when SLS_DEBUG is set? The debug statements we added only print if that is set. Also, would you mind listing your permissions? It would really help trying to find the root cause here, as I'm not able to reproduce the bug you've found.

It's interesting you state that sls create_domain also fails. Was that command working with previous versions?

FluxAugur commented 5 years ago

@captainsidd This is what I'm getting (with SLS_DEBUG=*). Note that 'api.my.domain' does not currently have a CNAME record (or any other) in Route53

Serverless: [AWS apigateway 404 0.479s 0 retries] getBasePathMappings({ domainName: 'api.my.domain' }) Serverless Domain Manager: NotFoundException: Invalid domain name identifier specified

Error --------------------------------------------------

Error: Unable to get BasePathMappings for api.my.domain.

 For debugging logs, run again after setting the "SLS_DEBUG=*" environment variable.

Stack Trace --------------------------------------------

Error: Error: Unable to get BasePathMappings for api.my.domain. at ServerlessCustomDomain. (/Users/lastson/Code/Repos/sandbox/serverless-api-gateway-domain/node_modules/serverless-domain-manager/dist/index.js:461:23) at Generator.throw () at rejected (/Users/lastson/Code/Repos/sandbox/serverless-api-gateway-domain/node_modules/serverless-domain-manager/dist/index.js:5:65) at processTicksAndRejections (internal/process/next_tick.js:81:5)

ryderjgillen commented 5 years ago

I continue to intermittently experience this issue. Logging into API gateway and deleting the base path mapping only, then retry, this seems to resolve the issue.

captainsidd commented 5 years ago

@FluxAugur have you run sls create_domain? If so, I'd be very curious as to why route53 records for your domain were not created and what logs showed up in SLS_DEBUG mode for create_domain

FluxAugur commented 5 years ago

@captainsidd I have. It fails. There are no CNAME records other than the domain validation record. Here's the debug info.

Serverless: Invoke create_domain Serverless: [AWS apigateway 404 0.612s 0 retries] getDomainName({ domainName: 'api.my.domain' }) Serverless Domain Manager: NotFoundException: Invalid domain name identifier specified Serverless: Selected specific certificateArn Serverless: [AWS apigateway 400 1.474s 0 retries] createDomainName({ certificateArn: '', domainName: 'api.my.domain', endpointConfiguration: { types: [ 'EDGE', [length]: 1 ] }, regionalCertificateArn: undefined }) Serverless Domain Manager: BadRequestException: One or more of the CNAMEs you provided are already associated with a different resource. (Service: AmazonCloudFront; Status Code: 409; Error Code: CNAMEAlreadyExists; Request ID: 04203862-5630-11e9-a106-ed88d4c45bcd)

Error --------------------------------------------------

Error: Failed to create custom domain api.my.domain

 For debugging logs, run again after setting the "SLS_DEBUG=*" environment variable.

Stack Trace --------------------------------------------

Error: Error: Failed to create custom domain api.my.domain

at ServerlessCustomDomain.<anonymous> (/Users/lastson/Code/Repos/sandbox/serverless-api-gateway-domain/node_modules/serverless-domain-manager/dist/index.js:335:15)
at Generator.throw (<anonymous>)
at rejected (/Users/lastson/Code/Repos/sandbox/serverless-api-gateway-domain/node_modules/serverless-domain-manager/dist/index.js:15:34)
at processTicksAndRejections (internal/process/next_tick.js:81:5)

FluxAugur commented 5 years ago

@captainsidd I just tried from a completely new project with the certificate still 'Issued' and not in use. No CloudFront distribution exists. Route53 hosted zone was new (with only 4 NS records and 1 SOA record. It succeeds. Here's all the relevant details.

# serverless.yml

service: products-service

provider:
  name: aws
  runtime: python3.6
  stage: dev
  region: us-east-1
  environment:
    SERVICE_NAME: ${self:service}

custom:
  customDomain:
    domainName: api.my.domain
    basePath: products
    stage: ${self:provider.stage}
    createRoute53Record: true

functions:
  hello:
    handler: handler.hello
    events:
      - http: GET hello

plugins:
  - serverless-domain-manager

Serverless: Invoke create_domain
Serverless: [AWS apigateway 404 0.168s 0 retries] getDomainName({ domainName: 'api.my.domain' })
Serverless Domain Manager: NotFoundException: Invalid domain name identifier specified
Serverless: [AWS acm 200 0.137s 0 retries] listCertificates({ CertificateStatuses: [ 'PENDING_VALIDATION', 'ISSUED', 'INACTIVE', [length]: 3 ] })
Serverless: [AWS apigateway 201 2.874s 0 retries] createDomainName({ certificateArn:
   '<valid certificate ARN>',
  domainName: 'api.my.domain',
  endpointConfiguration: { types: [ 'EDGE', [length]: 1 ] },
  regionalCertificateArn: undefined })
Serverless: [AWS route53 200 0.152s 0 retries] listHostedZones({})
Serverless: [AWS route53 200 0.139s 0 retries] changeResourceRecordSets({ ChangeBatch:
   { Changes:
      [ { Action: 'UPSERT',
          ResourceRecordSet:
           { AliasTarget:
              { DNSName: '<valid distribution>.cloudfront.net',
                EvaluateTargetHealth: false,
                HostedZoneId: 'Z2FDTNDATAQYW2' },
             Name: 'api.my.domain',
             Type: 'A' } },
        { Action: 'UPSERT',
          ResourceRecordSet:
           { AliasTarget:
              { DNSName: '<valid distribution>.cloudfront.net',
                EvaluateTargetHealth: false,
                HostedZoneId: 'Z2FDTNDATAQYW2' },
             Name: 'api.my.domain',
             Type: 'AAAA' } },
        [length]: 2 ],
     Comment: 'Record created by serverless-domain-manager' },
  HostedZoneId: '<valid hosted zone ID' })
Serverless: Custom domain api.my.domain was created.
            New domains may take up to 40 minutes to be initialized.

After create_domain successfully completes, sls deploy completes with no issues.

Based on the above, I'm assuming that you can't use this plugin on a hosted zone that already has apex alias records (A ALIAS/AAAA ALIAS)?

The instance where create_domain fails is where I already have an existing hosted zone with apex records (e.g. my.domain) with a certificate for my.domain and a CloudFront distribution listing my.domain as the only CNAME, and I'm trying to add api.my.domain with this plugin.

Is this not the correct usage? Why would this create A/AAAA records instead of just CNAME pointing api.my.domain to the .cloudfront.net? Since it creates apex alias records, does that mean that I would have to have api.my.domain in a separate hosted zone?

I hope the above makes some modicum of sense. And thanks in advance for your help.

FluxAugur commented 5 years ago

I got it working. Apparently, there was some issue with the certificate. When I deleted it and created a new one, like magic, everything worked. I still thought you could only have one apex record per hosted zone (DNS namespace). Or doesn't this include apex alias records?

mattdell commented 5 years ago

I'm still having this problem.

service: my-serverless

provider:
  name: aws
  runtime: nodejs8.10
  stage: ${env:STAGE}
  region: eu-west-1
  apiGateway:
    minimumCompressionSize: 1024

plugins:
  - serverless-dotenv-plugin
  - serverless-nextjs-plugin
  - serverless-domain-manager
  - serverless-offline

custom:
  serverless-nextjs:
    nextConfigDir: ./
  customDomain:
    domainName: serverless-${env:STAGE}.mywebsite.com
    certificateName: 'mywebsite.com'
    createRoute53Record: true

package:
  # exclude everything
  # page handlers are automatically included by the plugin
  exclude:
    - ./**

Whenever I create a new stack, I get

Error: Unable to get BasePathMappings for serverless-dev.mywebsite.com

felimartina commented 5 years ago

Make sure you run sls create_domain before running sls deploy. That did the trick for me.

mattdell commented 5 years ago

Yeah, I just started doing that and it seems to fix the problem.

Is that how it's meant to work? I didn't see anything about that in the docs.

Edit: Just re-read the docs. It's there. 🤦‍♂️

obernardocosta commented 5 years ago

This erro happens when I tried to delete all environment (i.e., I delete cf stack) and tried deploy again with sls deploy. @felimartina solutions works!

seba-i commented 5 years ago

I found that I had to set endpointType (default is 'edge', I needed 'regional') to solve this error.

sscarduzio commented 4 years ago

It's very confusing that I had to create first the domain, and then deploy. In my opinion this should be at least logged in the output of the failing create_domain.

abdodeve commented 4 years ago

Me to i had the same problem. The problem was that i used another AWS account. Be sure you use your right aws account . Verify it here: C:/Users/USER1/.aws/credentials

cmanzur commented 4 years ago

You should add endpointType, like this:

    domainName: dev.domain.com
    basePath: ''
    certificateName: '*.domain.com'
    stage: ${self:provider.stage}
    createRoute53Record: true
    endpointType: regional

Then: sls create_domain sls deploy -v

thiagomarini commented 4 years ago

Had a bit of a hard time with this error, I beleieve it was because I changed the region from us-east-1 to eu-west-1. Althought I had run sls remove on us-east-1 I think I forgot to to run sls delete_domain as well.

This sequence worked for me: sls delete_domain sls create_domain sls deploy

igoralves1 commented 4 years ago

I am getting the same issue after change the region from us-east-1 to us-west-2. Now if I try sls create_domain --stage mysatge I am getting Failed to create custom domain ..... If I run sls delete_domain --stage devig I am getting Serverless: Unable to delete custom domain .... If I try to deploy this is what I get Unable to get BasePathMappings for ... and cant deploy.

hoang-innomize commented 4 years ago

We turn on autoDomain flag, so the create_domain should be executed as a part of sls deploy command per documentation. Not sure why it doesn't work as expected.

bryan-hunter commented 4 years ago

@hoang-innomizetech autoDomain setting was just released yesterday. Do you have any more details about what is failing for you?

hoang-innomize commented 4 years ago

I am getting below error Serverless Domain Manager: NotFoundException: Invalid domain name identifier specified

Then I have to create domain before deploying sls service

47arathore commented 4 years ago

Note. You need to have the domain name or its parent as a hosted zone in the same AWS account where you have your API gateway. Then use the command sls create_domain and once it is initialised (it may take up to 40 mins), then run sls deploy.

fc-udaysolanki commented 3 years ago

So, problem with us was , we have not used small-case for domain name. I changed it to small-case and now it is working fine

amplify-education / serverless-domain-manager

Error: Unable to get BasePathMappings for domain.com #210