OC CLI: Environment variable with a DOT in them.

[gt-tech]

I am creating a new application using "oc new-app" CLI command and need to pass environment variables, one of my environment variable happens to be of the format of a.b.c.d=value, please see exact command below.

I get an error:

error: environment variables must be of the form key=value: spring.cloud.consul.host=http://192.168.18.132
error: environment variables must be of the form key=value: spring.cloud.consul.port=18500

Version

oc version
oc v1.2.0-rc1
kubernetes v1.2.0-36-g4a3f9c5

Steps To Reproduce

1. Create any application using "oc new-app" and try to pass an environment variable that looks like spring.cloud.consul.host=http://192.168.18.132

My sample command was -
oc new-app openshift/openshift/wildfly-100-centos7~https://codehub.myorg.com/my-lab/ose-sample-idp.git#master -e env=dev -e tenant=oid -e "spring.cloud.consul.host"="http://192.168.18.132" -e "spring.cloud.consul.port"="18500" -e MAVEN_ARGS="-s -f boot-war-pom.xml -am -pl api clean package install -Popenshift -DskipTests -B"

Current Result

Error on environment variables with DOT in them

Expected Result

With enclosed in quotes, it should work?

[mfojtik]

@smarterclayton this is pretty easy to fix in our CLI, but will require fix in k8s to allow dots in env vars. Do you have any other concerns about this except the variables wont export in bash? Also docker run -e foo.test=bar works, so I don't see the reason it should not work in k8s.

[christian-posta]

Can't you just use FOO_BAR_COFFEE ? ditch the '.' characters since http://stackoverflow.com/questions/2821043/allowed-characters-in-linux-environment-variable-names

[bparees]

closing since periods are not valid in env variable identifiers, at least in bash.

[smarterclayton]

Hold on - just because they're not in bash doesn't mean they aren't valid to pass to the container.

[smarterclayton]

We distinguish between Posix and Bash - it should be valid to define env vars with periods in them via the container interfaces in Kube and the OpenShift APIs. Some contexts we may not allow them to be set, but only if the only interface to them is bash.

[bparees]

@smarterclayton well k8s doesn't allow it, so unless we're going to get k8s to allow them, we can't fix this.

[bparees]

spec.template.spec.containers[0].env[0].name: Invalid value: "FOO.BAR": must be a C identifier (matching regex [A-Za-z_][A-Za-z0-9_]*): e.g. "my_name" or "MyName"

[smarterclayton]

Probably something we'll have to fix in Kube.

On Thu, Aug 4, 2016 at 11:38 PM, Ben Parees [email protected]
wrote:

spec.template.spec.containers[0].env[0].name: Invalid value: "FOO.BAR": must be a C identifier (matching regex [A-Za-z_][A-Za-z0-9_]*): e.g. "my_name" or "MyName"

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#8771 (comment),
or mute the thread
https:/notifications/unsubscribe-auth/ABG_pyUd-hsbekddUcptGoNOGSSdb81Bks5qcrArgaJpZM4IY1Rt
.

[smarterclayton]

Please spawn a Kube issue.

On Thu, Aug 4, 2016 at 11:40 PM, Clayton Coleman [email protected]
wrote:

Probably something we'll have to fix in Kube.

On Thu, Aug 4, 2016 at 11:38 PM, Ben Parees [email protected]
wrote:

spec.template.spec.containers[0].env[0].name: Invalid value: "FOO.BAR": must be a C identifier (matching regex [A-Za-z_][A-Za-z0-9_]*): e.g. "my_name" or "MyName"

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#8771 (comment),
or mute the thread
https:/notifications/unsubscribe-auth/ABG_pyUd-hsbekddUcptGoNOGSSdb81Bks5qcrArgaJpZM4IY1Rt
.

[bparees]

kubernetes/kubernetes#30120

[maiconbaumx]

Any news about this issue? I'm having the same problem here.. I want to deploy an Apache Zuul from NetflixOSS, but, it appears to need a env variable with "dot", like this: origin.zuul.client.DeploymentContextBasedVipAddresses=HELLO-NETFLIX-OSS

https:/Netflix-Skunkworks/zerotodocker/wiki/Running-the-Images

[bparees]

it's blocked by the upstream issue i linked above: kubernetes/kubernetes#30120

i haven't seen any movement on that issue.

[xiaods]

it'll only be available in 1.8.

[openshift-bot]

Issues go stale after 90d of inactivity.

Mark the issue as fresh by commenting /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.
Exclude this issue from closing by commenting /lifecycle frozen.

If this issue is safe to close now please do so with /close.

/lifecycle stale

[bparees]

looks like this is resolved so we might actually be able to fix this now: kubernetes/kubernetes#2707
/remove-lifecycle stale
/lifecycle frozen