Get hands-on experience with 20+ free Google Cloud products and $300 in free credit for new customers.

Trying to create a Ray on Vertex ai cluster and failing to do so on my project

I am following steps mentioned over here : https://cloud.google.com/vertex-ai/docs/open-source/ray-on-vertex-ai/create-cluster?hl=en and trying to create a cluster ai and it is always showing unhealthy status. I am not sure what is going wrong since its literally what the doc says. I have enabled the apis, added the networking part and all and even tried this with 3 different projects but facing same thing everywhere.

Would appreciate the tiniest nudge possible on this. I am a newbie in GCP but I followed the same steps in my previous account and it worked. The problem is my previous account is no more and need to use a new account to set this up.

Cluster config shown unhealtyCluster config shown unhealtydefault network configdefault network config

0 1 886
1 REPLY 1

I understand that you're trying to create a cluster on Google Vertex AI following the documentation, but the cluster is consistently showing an unhealthy status. Make sure you have enabled the necessary APIs and services in your GCP project.

Ensure that your firewall rules allow necessary traffic to and from the cluster. If you are using custom VPC networks, verify that the firewall rules are correctly configured to allow the required traffic. Review the node pool settings to make sure they are appropriate for your workload. Ensure that the machine types and the number of nodes are set to values that fit your use case. Make sure that the versions of the software components you are using (e.g., Ray, Kubernetes, Vertex AI) are compatible with each other.

If you've checked all these points and are still facing issues, I recommend looking at the error messages or logs you're receiving when creating the cluster. They may provide more specific details about what's going wrong. Additionally, consider reaching out to Google Cloud support for more personalized assistance, as they can access your project directly to help diagnose the issue.