https://www.gooddata.com/ logo
Join Slack
Powered by
Hi everybody, I am still stuck with deploying <Goo...
# gooddata-cn
r
Hi everybody, I am still stuck with deploying GoodData.cn in K8S on GKE. I have successfully deployed nginx and pulsar while for goodData, some pods aren't starting correctly. Those pods are : • gooddata-gooddata-cn-afm-exec-api • gooddata-gooddata-cn-api-gateway • gooddata-gooddata-cn-auth-service • gooddata-gooddata-cn-calcique • gooddata-gooddata-cn-scan-mode • gooddata-gooddata-cn-sql-executor All of them are stuck in pending state. Any idea about this ?
b
Hello Reda, there are several reasons why the pods might be unable to start. 
kubectl -n <namespace> describe pod <pod-name>
might contain some useful information as to why.
r
example :
Copy code
kubectl -n gooddata  describe pod gooddata-gooddata-cn-sql-executor-6695fb94db-9f8g4
give this
Copy code
Events:
  Type     Reason             Age                     From                Message
  ----     ------             ----                    ----                -------
  Warning  FailedScheduling   3m45s (x1011 over 22h)  default-scheduler   0/6 nodes are available: 6 Insufficient cpu.
  Normal   NotTriggerScaleUp  3m44s (x7921 over 22h)  cluster-autoscaler  pod didn't trigger scale-up:
is it only matter of cpu and I should add additional nodes. @Robert Moucha told me that my current capacity : 6x e2-medium, it's 12vCPU and 24GiB RAM should be ok
fyi, it's a dedicated cluster for good data$
b
it really looks like a sizing issue - please try checking the 
requests
and 
limits
of the pods you're deploying in 
values.yaml
of the helm chart. If there are any that define high 
requests
, they might not fit on the nodes you have.
it might also help to tune overcommit settings for your k8s
could you please provide the output of 
kubectl describe nodes
?
r
sure
here it is
image.png
b
this node looks ok, it has some spare capacity. What about the others?
r
image.png
image.png
b
From these I don't really see what the issue could be. Could you please send the output of 
kupectl get nodes -o json | jq '.items[].status | .capacity, .allocatable'
? This will only contain the sizes of the individual nodes
r
replacing e2-medium machine type by e2-standard fixed the issue
👍 1
all pods are running
I've also bind my subdomain to cluster public ip
and now, it should work right ?
m
Hi Reda, let's assume it is configured properly 🙂 You will definitely find out when trying to create an organization when clusterIssuer you have just created will be used - https://www.gooddata.com/developers/cloud-native/doc/2.1/manage-deployment/set-up-organizations/setup-organizations/
r
Thanks Matej
but I shouldn't receive 404 when I am trying to access my site
my yaml custom values :
Copy code
license:
    key: "myKey"
    
  dex:
    # -- base context path prefix used by Dex to serve its resources
    uriPrefix: /dex
    ingress:
      # -- hostname where the application will have its authentication Endpoint (Dex). It will be used for
      # organizations without their own external Identity Provider.
      authHost: '<http://mydomain.io|mydomain.io>'
      tls:
        # -- If you have pre-existing secret with your own certificate and key, put its name here.
        # Also, if you want cert-manager, set to some Secret name where TLS certificate and key will be stored.
        # Note that dex.ingress.authHost is required when enabling TLS.
        # If you're deploying to AWS, you may prefer TLS termination on AWS ELB and keep this value empty.
        authSecretName: 'gooddata-cn-auth-tls'
      # -- Custom annotations that will be added to eery Ingress object created by this chart, e.g.
      # or using namespace-specific Issuer:
      # <http://cert-manager.io/issuer|cert-manager.io/issuer>: local-ca-issuer
      annotations:
         <http://cert-manager.io/cluster-issuer|cert-manager.io/cluster-issuer>: letsencrypt-production

  ingress:
    # -- Class of the Ingress controller used for this deployment
    ingressClassName: nginx
I voluntarily changed the licence key and auth domain
can you tell me what's missing here ?
m
you got 404 from where? Did you managed to create the organization?
r
not yet
is it a prerequisites organization creation ?
because when I see 404, it looks like something wrong with the ingress no ?
Thought I can do so, handle organization
I got 404 when trying to access mydomain.io
the site I've create A record for in my DNS
m
ingress controller returns 404 because by default you have no configured ingress objects
if you create an organization then the ingress object for organization will be created - and you can access it
r
hat'sx great
let me test
m
👍
r
Hi Matej
it finally worked
image.png
m
perfect!
Open in Slack
PreviousNext
Powered by