PE code deployment fails with undefined method when external modules declared in Puppetfile

asked 2016-10-29 12:15:07 -0500

swarly gravatar image

I've created a basic Puppet Enterprise setup for testing. It has a puppet_control repo (using the Puppet-supplied example), uses Code Manager and only pulls in a handful of modules from the Forge and has a few manifests.

However, whenever I run puppet-code deploy --all --wait, I get the following error:

[{"environment":"production","id":5,"status":"failed","error":{"kind":"puppetlabs.code-manager/deploy-failure","details":{"corrected-env-name":"production"},"msg":"Errors while deploying environment 'production' (exit code: 1):\nERROR\t -> undefined method `each' for #<String:0x00000001179d68>\n"}}]

(Running puppet-code deploy with a -l debug or -l trace does not produce anything more in-depth on that error.

What's interesting is that if I remove all of the module lines from Puppetfile, i.e. don't import any modules, then the deployment works and the code is synced from /etc/puppetlabs/code-staging/ to /etc/puppetlabs/code/.

The only wrench in this whole setup that may be affecting this is that the Puppet Master/Server has to go through a web proxy to get out to the Internet. However, this works. OTOH, I have no idea if Code Manager isn't following what it sees in r10k.yaml.

Any ideas would be appreciated.

edit retag flag offensive close merge delete

Comments

I manage a few PE deployments behind web proxies that are running Code Manager. Assuming that you've got puppet_enterprise::profile::master::r10k_proxy configured and aren't running into the RK-229 issue. Can you post the Puppetfile that fails?

aharden gravatar imageaharden ( 2016-10-31 06:20:48 -0500 )edit