Exit code 0 even though deployment fails

[johnsgp]

I'm using serverless 0.5.5.

When I run serverless event deploy before deploying the lambda function, I get something like this:

Serverless: Deploying events in "dev" to the following regions: eu-west-1
Serverless: Failed to deploy events in "dev" to the following regions:
Serverless: eu-west-1 ------------------------
Serverless:   schedule: Cannot find alias arn: arn:aws:lambda:eu-west-1:.....
Serverless:
Serverless: Run this again with --debug to get more error information...

The process returns exit code 0 which I then cannot detect as an error with a continuous integration tool.

There are similar problems with deployment of functions and endpoints, e.g., when the _meta/_tmp folder is not writeable. The common feature is the last line, prompting to run again with --debug.

Is there an option to make such failures into return exit code 1?

[stevecaldwell77]

I'm running into the same issue. Came in this morning to find my lambda functions hadn't been executing overnight, because deploying their scheduled event failed silently in travis.

As a side note, an event name longer than 64 chars seems like a common-ish use case with the way serverless builds them up.

Serverless: Deploying events in "travis" to the following regions: us-east-1
Serverless: Failed to deploy events in "travis" to the following regions:
Serverless: us-east-1 ------------------------
Serverless:   myLongEventName012345: 1 validation error detected: Value 'my_project_name-my_longish_function_name-myLongEventName012345-travis' at 'name' failed to satisfy constraint: Member must have length less than or equal to 64
Serverless:
Serverless: Run this again with --debug to get more error information...

[csterwa]

I'm getting the same with failed deployment of endpoints:

Serverless: Failed to deploy endpoints in "dev" to the following regions:  
Serverless: us-east-1 ------------------------  

[flomotlik]

@johnsgp this is resolved in Serverless V1. The system exits properly with 1 if the deployment fails so I'll close this issue. Thanks for reporting.

[RickvdP]

Looks like this is still happening, did I miss anything?

[Container] 2018/05/27 14:26:48 Running command serverless deploy --stage cicd | tee deploy.out
Serverless: Packaging service...
Serverless: Excluding development dependencies...
Serverless: Tracing ENABLED for function "template-service-cicd-hello"
 
  Serverless Error ---------------------------------------
 
  ServerlessError: User: arn:aws:sts::1234:assumed-role/code-build-template-service-service-role/AWSCodeBuild-abc is not authorized to perform: cloudformation:DescribeStacks on resource: arn:aws:cloudformation:eu-west-1:1234:stack/template-service-cicd/*
 
  Get Support --------------------------------------------
     Docs:          docs.serverless.com
     Bugs:          github.com/serverless/serverless/issues
     Issues:        forum.serverless.com
 
  Your Environment Information -----------------------------
     OS:                     linux
     Node Version:           6.3.1
     Serverless Version:     1.27.3
 

[Container] 2018/05/27 14:26:52 Phase complete: BUILD Success: true

Amazon CodeBuild still marks the output as Success: true. I just verified that outputting exit 1 actually translates into Success: false too.

Any suggestions?

[traviswimer]

I'm seeing the same problem.

[traviswimer]

Actually, nevermind. Serverless does seem to be using the correct exit code.

I think I was getting a 0 exit code because I was running it as a script in my package.json (e.g. npm run serverless_thing). Apparently the exit codes of the commands within an npm script do not get passed through.

@RickvdP , I would bet your problem is caused by piping into tee deploy.out, since that part would presumably succeed.

[abmohan]

@traviswimer - thanks for the update. I encountered the same problem

[kaitlynbrown]

I'm encountering the same problem
edit: turned out to be because of an issue in serverless-offline fixed by dherault/serverless-offline#767