https://pulumi.com logo
d

damp-book-35965

04/02/2019, 10:11 PM
I'm getting this error:
Copy code
Diagnostics:
  pulumi:pulumi:Stack (xxx):
    error: update failed
 
  kubernetes:core:Service (kube-system/prometheus-operator-prometheus):
    error: Plan apply failed: 2 errors occurred:
    
    * Resource operation was cancelled for 'prometheus-operator-prometheus'
    * Service was not allocated an IP address; does your cloud provider support this?
What's really happening ? There are enough IPs in the VPC to provision
c

creamy-potato-29402

04/02/2019, 10:14 PM
@damp-book-35965 you cancelled the operation
maybe we should not display the second error when you cancel, but it just means your service did not complete initializing.
d

damp-book-35965

04/02/2019, 10:15 PM
Nope..this is repeatedly happening since 2-3 times so now I'm not waiting for it to timeout
c

creamy-potato-29402

04/02/2019, 10:16 PM
when we say the operation was “cancelled” we mean that the kubernetes provider was sent the
Cancel
RPC call.
d

damp-book-35965

04/02/2019, 10:17 PM
otherwise stuck on this:
Copy code
kubernetes:helm.sh:Chart      prometheus-operator                                      
 ~      └─ kubernetes:core:Service    kube-system/prometheus-operator-prometheus  updating..     [2/3] Attempting to allocate IP address to Service
c

creamy-potato-29402

04/02/2019, 10:17 PM
When I wrote this, there was exactly one place where this was called — the ctrl-c handler.
@damp-book-35965 and when you wait, it times out?
d

damp-book-35965

04/02/2019, 10:17 PM
Yes..
c

creamy-potato-29402

04/02/2019, 10:17 PM
what type is your service.
d

damp-book-35965

04/02/2019, 10:18 PM
LoadBalancer
c

creamy-potato-29402

04/02/2019, 10:22 PM
@damp-book-35965 when it times out, does it actually have an IP attached?
d

damp-book-35965

04/02/2019, 10:31 PM
Nope. it doesnt
c

creamy-potato-29402

04/02/2019, 10:31 PM
sounds like it did not finish initializing then
the error was truthful. 🙂
d

damp-book-35965

04/02/2019, 10:32 PM
I think something else is going on..So are you saying we should increase timeout or something :?
c

creamy-potato-29402

04/02/2019, 10:32 PM
no
d

damp-book-35965

04/02/2019, 10:32 PM
It's not telling me the reason for not getting an IP .
c

creamy-potato-29402

04/02/2019, 10:33 PM
When I ask whether the service has an IP address attached, I mean when you
kubectl get
it, is there an IP address reported in the
.status
field.
it’s not telling you the reason, because Kubernetes does not know the reason.
d

damp-book-35965

04/02/2019, 10:33 PM
Let me check that
c

creamy-potato-29402

04/02/2019, 10:34 PM
when you say it wasn’t allocated an IP, why did you say no?
d

damp-book-35965

04/02/2019, 10:34 PM
Because it's a load balancer type and there is no load balancer spun up on AWS
c

creamy-potato-29402

04/02/2019, 10:35 PM
ok, well, in any event, in general we’ll have no idea why the service does not create a LB.
d

damp-book-35965

04/02/2019, 10:36 PM
When I do kubectl get svc it shows external IP as pending
c

creamy-potato-29402

04/02/2019, 10:36 PM
it should. the service controller should be talking to AWS and spinning one up. If it doesn’t work, it’s up to the service controller to report the error through the status field
right, so that sounds like some kind of problem with AWS or the service controller.
d

damp-book-35965

04/02/2019, 10:36 PM
Yes..
c

creamy-potato-29402

04/02/2019, 10:37 PM
but we don’t know what it is.
can you show me the output of the
.status
field with
-o yaml
d

damp-book-35965

04/02/2019, 10:37 PM
Yup..I was wondering what's the logic from the pulumi error message
Copy code
status:
  loadBalancer: {}
c

creamy-potato-29402

04/02/2019, 10:38 PM
yeah
ok so there was no error to report, which is not great.
d

damp-book-35965

04/02/2019, 10:41 PM
Found something else with kubectl describe events
Which gave me my answer: we reached our limits for ELBs 😥
c

creamy-potato-29402

04/02/2019, 10:43 PM
@damp-book-35965 ah, we should be including that information in the error
cc @gorgeous-egg-16927
👍 1
this is part of the “events” workstream
d

damp-book-35965

04/02/2019, 10:43 PM
Yes
w

white-balloon-205

04/03/2019, 2:54 AM
Specifically this is part of the work tracked in https://github.com/pulumi/pulumi-kubernetes/issues/133.