Stuck on Service Mesh Fundamentals Lab 4.1 (ingress ctrl/Ambassador install)

Irvnet

Hey!
I'm working on Lab 4.1 for the Service Mesh Fundamentals class. Has anyone been able to get Ambassador installed successfully? I'm a bit confused about how to get it running... seems there are changes with Ambassador since the last release of the class. When I look at the instructions on ambassador's website its not clear which install instructions you should follow... there are several including:

• https://github.com/datawire/ambassador-operator/blob/master/docs/install.md
• https://www.getambassador.io/docs/edge-stack/latest/topics/install/yaml-install/
• https://www.getambassador.io/docs/emissary/2.1/tutorials/getting-started/

Lab 1.1 suggests k8s version was kindest/node:v1.18.2 was used... tough to match the ambassador version since the provided paths in the lab doc point to "latest".

i'm using:

• AWS ec2: T3.xlarge / ubuntu 18.04
• helm v3.8.1
• kind version: v0.12 w/ kindest/node:v1.19.16
• docker community v20.10.14
• Ambassador: edge stack v 2.2.2 ( https://app.getambassador.io/yaml/edge-stack/2.2.2/aes.yaml )

If I use the current Ambassador it makes the service of type LoadBalancer (and puts it in a pending state). Can I choose what kind of LB ambassador makes or do I need to install a load balancer?

Not sure how to proceed here...

Irvnet

Ok... so i ended up using metallb [ https://metallb.universe.tf/ ] ... That at least got me an external IP on the ambassador load balancer. Didn't quite solve my problem, but I have to do a bit more debugging. either I didn't reserve the right ip addresses (I just took the defaults) or i have to review the ambassador mapping... or just have to reinstall it all from the beginning again.

The instructions direct you to install each component across all 3 clusters each time. I'm kinda doing the opposite... trying to get at least one cluster working first instead of doing installs on 3 broken ones

Anyone have any better luck than I have?

R-

Irvnet

Hey... kept plugging...
The example kind provides [ https://kind.sigs.k8s.io/docs/user/ingress/#create-cluster ] seems to work pretty consistently for getting Ambassador installed in a kind cluster (and getting the their provided sample app to respond).

Heading out to a blockchain meetup so I'll get back to it tomorrow... next:

• I'll try the cluster config and ambassador install from the kind docs, then install the sample app from the labs to see if that works.
• take a shot at deploying consul as described in the labs to see how that goes.

What I'd really like are some tips to debug/validate the network traffic as it runs thru the ingress controller going to the service... (i.e. network diag commands independent any service mesh commands). if the trail is broken between the client and back end pod, how do I find out who's not talking to who?

Let me know what you find...

R-