K
Q

GKE cluster fails to create instances

11/5/2018

Creating a gcloud kubernetes cluster from a script and the cluster is takes forever to create, just ends up failing after 35 minutes.

Command:

gcloud container clusters create 148374ed-92b0-4088-9623-c22c5aee3 \
    --num-nodes 3 \
    --enable-autorepair \
    --cluster-version 1.11.2-gke.9 \
    --scopes storage-ro \
    --zone us-central1-a

The error are not clear, looks like some kind of buffer overflow internal to gcloud.

Deploy error: Not all instances running in IGM after 35m6.391174155s. Expect 3.

Current errors: [INTERNAL_ERROR]: Instance 'gke-148374ed-92b0-default-pool-66d3729f-6mw3' creation failed: Code: '-2097338327842179396' - ; Instance 'gke-148374ed-92b0-default-pool-66d3729f-qwpd' creation failed: Code: '-2097338327842179396' - ; .

Any ideas for debugging this?

-- Graham
google-cloud-platform
google-kubernetes-engine

Similar Questions

Store Vault secrets in S3 bucket
Couldn't initialize a Kubernetes cluster with Vagrant
How to get the audit, file beat and metric beat logs in kubernetes cluster using EFK stack?
I can't get through to Hadoop server from Hadoop client
Kubernetes pods cannot reach external ip addresses, but can access any other internal ip address inside the cluster. such as pod's ip, service's ip

1 Answer

11/5/2018

I've been facing similar issue while creating a cluster for the past 3 hours. A ticket has already been raised and GCP engineering team is working on the fix.

For status updates on the ticket, visit https://status.cloud.google.com/incident/compute/18012

-- Rushil Basappa
Source: StackOverflow