Gateway receives no APIs from Dashboard


#1

I am trying out Tyk Pro so I deployed a gateway, the dashboard and pump to k8s by making some small modifications to the official helm chart. I was able to log in to the dashboard and I created an API but when I look at the logs of the gateway I see that it doesn’t receive the API from the gateway even though the configuration is reloaded right as I update my API:

time="Feb 21 14:00:20" level=info msg="Reloading endpoints"
time="Feb 21 14:00:20" level=info msg="Reload queued"
time="Feb 21 14:00:20" level=info msg="reload: initiating"
time="Feb 21 14:00:20" level=info msg="Loading policies"
time="Feb 21 14:00:20" level=info msg="Using Policies from Dashboard Service"
time="Feb 21 14:00:20" level=info msg="Mutex lock acquired... calling"
time="Feb 21 14:00:20" level=info msg="Calling dashboard service for policy list"
time="Feb 21 14:00:20" level=info msg="Processing policy list"
time="Feb 21 14:00:20" level=info msg="Policies found (0 total):"
time="Feb 21 14:00:20" level=info msg="Detected 0 APIs"
time="Feb 21 14:00:20" level=warning msg="No API Definitions found, not reloading"
time="Feb 21 14:00:20" level=info msg="reload: complete"
time="Feb 21 14:00:20" level=info msg="Initiating coprocess reload"

So it looks like the communication between the Dashboard and the Gateway is fine, but the Dashboard is not sending my API to the gateway? The API is enabled for sure.


#2

I was able to figure out that this was happening due to having gateway sharding turned on in the default configuration. To the authors of tyk: I’d recommend turning this off by default as it results in a very non-intuitive failure for new users.


#3

Oh thats indeed confusing!

Can you pls give link to configuration which had sharding enabled?

Thank you!


#4

As I mentioned in the first post I had to make some adjustments to the helm chart to get up and running as it’s a bit all over the place. I am hoping to submit a PR with my changes to make it easier to get up and running with the helm chart as I imagine it’s the way for most people to get up and running on k8s.


#5

If we make that change to the default config it breaks the use of the gateway as an ingress.

However a seperate helm chart that is just for isolated k8s deployments make sense - can yoiu share the one you created @andor44