Hi Hemant,

As Matt mentioned, with recent versions of the armada code base, if you install the armada chart (inside the cluster), it creates a tiller sidecar by default (can be disabled though) and connects to that over localhost.

However, since you are installing armada outside of the cluster, I assume you are likely using the armada CLI? In that case, by default it will look up the cluster ip for the tiller pod and use that.
However, since you are outside the cluster that won't be accessible. So you will then need to expose your tiller outside the cluster and then use it's address as --tiller-host on the command line, or tiller_host in your /etc/armada/armada.conf, like this:

[DEFAULT]
<...other content...>
tiller_host=tiller.my.cluster

Thanks,
Sean


From: MCEUEN, MATT
Sent: Wednesday, July 31, 2019 9:03 AM
To: Hemant Sonawane; airship-discuss@lists.airshipit.org
Subject: Re: [Airship-discuss] Failed to connect to remote host: FD Shutdown

***Security Advisory: This Message Originated Outside of AT&T ***
Reference http://cso.att.com/EmailSecurity/IDSP.html for more information.

Hi Hemant,

 

I see a timeout buried in that error log.  A couple questions to get context: 

 

Are you using Armada standalone to talk to an existing Kubernetes cluster, or are you using the full Airship stack?


If the former, which version/ref of Armada and tiller are you using; if the latter, which version of Treasuremap, Airship-in-a-bottle (etc) are you using as a reference?

 

In current versions, Armada provides its own tiller as a sidecar, and connects to it via localhost, although this behavior is overridable and you can alternately “bring your own tiller”.

 

Thanks,

Matt

 

From: Hemant Sonawane <hemant.sonawane@itera.io>
Sent: Wednesday, July 31, 2019 3:25 AM
To: airship-discuss@lists.airshipit.org
Subject: [Airship-discuss] Failed to connect to remote host: FD Shutdown

 

Hi All,

 

Deploying OpenStack from outside of Kubernetes cluster causing me some issue and I am using armada for the deployment. I am getting connection error related to "target_address" and honestly I really don't understand why is it like that. Also Is there any way to change the target_address? as now its pointing to tiller(helm) ClusterIP address. Basically my goal is to have contact with OpenStack-helm to be able to deploy OpenStack from outside the cluster.  For better understanding of my issue please find attached logs and let me know if there is any solution to resolve it.

 

--

Thanks and Regards,

 

Hemant Sonawane

Image removed by sender.