Prisma1: Unable to load Prisma config: mapping values are not allowed here

[aitchkhan]

Describe the bug
When trying to deploy prisma to AWS ECS + EC2, on setting the env variable PRISMA_CONFIG.

The value of the variable looks like this:
port: 4466 PRISMA_MANAGEMENT_API_SECRET: sesd-secret-api databases: default: connector: mysql host: aaa.ssss.us-east-1.rds.amazonaws.com port: 3306 user: foobaz password: XXX migrations: true.

Note: I am not using AWS Fargate.

Getting the following error:

Exception in thread "main" java.lang.RuntimeException: Unable to load Prisma config: mapping values are not allowed here
in 'string', line 1, column 40:
... 466 PRISMA_MANAGEMENT_API_SECRET: khareta-secret-api databases: ...

Setting the env variable in AWS task definition:

Expected behavior
Prisma containers should be deployed.

Versions (please complete the following information):

• Connector: [MySQL]
• Prisma Server: [1.13.0]
• prisma CLI: (not needed)
• OS: [ Ubuntu 16.04]
  
  
  
  • Docker image: [prismagraphql/prisma:1.13.0]
  
  

[pantharshit00]

You will need to pass it like so

| \n        port: 4466\n        managementApiSecret: sesd-secret-api\n        databases:\n          default:\n            connector: mysql\n            host: aaa.ssss.us-east-1.rds.amazonaws.com \n            port: 3306\n            database: database\n            user: foo\n            password: bar\n            migrations: true\n

We also support a PRISMA_CONFIG_PATH if this feels ugly to you. Here is an example: https://www.prisma.io/tutorials/deploy-prisma-to-dokku-ct15/#create-a-new-project-folder

[aitchkhan]

I tried the first option:
2019-02-13 08:37:56 at scala.Function0.apply$mcV$sp$(Function0.scala:34) 2019-02-13 08:37:56 at scala.runtime.AbstractFunction0.apply$mcV$sp(AbstractFunction0.scala:12) 2019-02-13 08:37:56 at scala.App.$anonfun$main$1$adapted(App.scala:76) 2019-02-13 08:37:56 at scala.collection.immutable.List.foreach(List.scala:389) 2019-02-13 08:37:56 at scala.App.main(App.scala:76) 2019-02-13 08:37:56 at scala.App.main$(App.scala:74) 2019-02-13 08:37:56 at com.prisma.local.PrismaLocalMain$.main(PrismaLocalMain.scala:12) 2019-02-13 08:37:56 at com.prisma.local.PrismaLocalMain.main(PrismaLocalMain.scala) 2019-02-13 08:37:56 at scala.sys.package$.error(package.scala:27) 2019-02-13 08:37:56 at com.prisma.config.ConfigLoader$.load(ConfigLoader.scala:40) 2019-02-13 08:37:56 at com.prisma.local.PrismaLocalDependencies.<init>(PrismaLocalDependencies.scala:40) 2019-02-13 08:37:56 at com.prisma.local.PrismaLocalMain$.delayedEndpoint$com$prisma$local$PrismaLocalMain$1(PrismaLocalMain.scala:15) 2019-02-13 08:37:56 at com.prisma.local.PrismaLocalMain$delayedInit$body.apply(PrismaLocalMain.scala:12) 2019-02-13 08:37:56 at scala.Function0.apply$mcV$sp(Function0.scala:34) 2019-02-13 08:37:56Exception in thread "main" java.lang.RuntimeException: Unable to load Prisma config: while scanning a block scalar 2019-02-13 08:37:56 in 'string', line 1, column 1: 2019-02-13 08:37:56 | \n port: 4466\n managementApiS ... 2019-02-13 08:37:56 ^ 2019-02-13 08:37:56expected a comment or a line break, but found \(92) 2019-02-13 08:37:56 in 'string', line 1, column 3: 2019-02-13 08:37:56 | \n port: 4466\n managementApiSec ... 2019-02-13 08:37:56 ^
Just getting a different error message.

I will give a shot to the other one...

[paulrostorp]

Any luck with this ?

[captDaylight]

I'm running into this same issue. My two cents, it's cumbersome to have one long env var like this.

[captDaylight]

@aitchkhan @pantharshit00 I've had the same exact trajectory, I've gotten both of these errors after modifying the variable string. Is there a standard way to format the PRISMA_CONFIG var that works on AWS?

[aitchkhan]

The only way I found was to use Edit Stack from AWS CloudFormation. You can update the variables there. Other than that, no luck.

[gabrielgian]

This one worked for me!

"port: 4466\ndatabases:\n default:\n connector: postgres\n host: postgres\n database: db\n port: 5432\n user: foo\n password: bar\n migrations: false\n schema: public\n"

You can use some online YAML to JSON converter.

[batmanhit]

@gabrielgian, your solution is still making an issue...
No way to achieve this on AWS console?

Has anyone solved this issue perfectly?

port: 4466\ndatabases:\n    default:\n        connector: mysql\n        host: mysql\n       port: 3306\n       user: prisma\n        password: password\n        migrations: true\n        rawAccess: true\n

[stefanfuchs]

So, I had no luck with any of the solutions presented above.
The only thing that worked for me was creating a custom image with a config.yml file inside, as presented here.
It works fine this way, but the prisma settings become hard coded in the image (though it's ok for my use case).

If anyone wants to try the same, I'll list the steps I've followed to make it work:

• Create a custom prisma image with a config.yml as described here. I'll post the full guide bellow, just in case the link gets broken some day:

• Create a new folder on your local machine and add these two files with the following content to it (customize them with your own configurations):

Dockerfile:

FROM prismagraphql/prisma:1.34.6
ARG PRISMA_CONFIG_PATH
ENV PRISMA_CONFIG_PATH prisma.yml
COPY config.yml prisma.yml
EXPOSE 4466

config.yml:

port: 4466
databases:
  default:
    connector: postgres
    host: dokku-postgres-prisma-server-db
    port: 5432
    user: postgres
    password: bd793d8a330715891698434697a0345b
    database: prisma

• Build the docker image and push it to Amazon ECR

• Use the image you just pushed instead of the original prisma image

• ???

• Profit