add rag kuberay and jupyterhub image #440
Conversation
|
/gcbrun |
|
/gcbrun |
1 similar comment
|
/gcbrun |
|
Going to try again as it failed due to trying to port fowarding rag-frontend which this PR shouldn't really be affecting |
|
/gcbrun |
chiayi
force-pushed
the
jupyterhub-rag
branch
3 times, most recently
from
cd639b4 to
8d3ff81
Compare
|
/gcbrun |
|
/gcbrun |
|
First attempt succeeded and the second attempt failed at calling webhook. |
|
/gcbrun |
* add rag kuberay and jupyterhub image (#440) * Rollback to previous image (#454) * Ray Webhook Support for Single-Host, Multi-Slice TPUs (#453) * Fix incorrect replicaIndex for single-host, multi replica * Fix single-host, multi-slice deletion logic * Update README & simplify workloads.tfvars for RAG (#445) * RAG marketplace updates (#456) * fix RAG marketplace changes --------- Co-authored-by: Chia-Yi Liang <chiayiliang327@gmail.com> Co-authored-by: zlq <zlq@google.com> Co-authored-by: ryanaoleary <113500783+ryanaoleary@users.noreply.github.com> Co-authored-by: imreddy13 <132504814+imreddy13@users.noreply.github.com>
| @@ -44,7 +44,8 @@ resource "helm_release" "ray-cluster" { | |||
| security_context = local.security_context | |||
| secret_name = var.db_secret_name | |||
| cloudsql_instance_connection_name = local.cloudsql_instance_connection_name | |||
| image_tag = var.enable_gpu ? "2.9.3-py310-gpu" : "2.9.3-py310" | |||
| image = var.use_custom_image ? "us-central1-docker.pkg.dev/ai-on-gke/rag-on-gke/ray-image" : "rayproject/ray" | |||
| image_tag = var.enable_gpu ? "2.9.3-py310-gpu" : var.use_custom_image ? "2.9.3-py310-gpu" : "2.9.3-py310" | |||
There was a problem hiding this comment.
@chiayi can we simplify this by also publishing 2.9.3-py310 to artifact registry? Even withput GPUs I think this would benefit
There was a problem hiding this comment.
Yes, I will work on creating an image for 2.9.3-py310 as well.
| @@ -121,6 +121,8 @@ resource "helm_release" "jupyterhub" { | |||
| gcs_bucket = var.gcs_bucket | |||
| k8s_service_account = var.workload_identity_service_account | |||
| ephemeral_storage = var.ephemeral_storage | |||
| notebook_image = "jupyter/tensorflow-notebook" | |||
There was a problem hiding this comment.
Was this meant to be updated with the artifact registry image?
|
|
||
| variable "use_custom_image" { | ||
| type = bool | ||
| description = "If running RAG, set this var to true to use custome image with pre-installed lib" |
There was a problem hiding this comment.
This description shouldn't mention RAG, it's a separate variable unrelated to RAG, but we'll set it to true in the RAG deployment
| @@ -228,6 +228,7 @@ module "kuberay-cluster" { | |||
| grafana_host = module.kuberay-monitoring.grafana_uri | |||
| disable_network_policy = var.disable_ray_cluster_network_policy | |||
| depends_on = [module.kuberay-operator] | |||
| use_custom_image = true | |||
There was a problem hiding this comment.
We shouldn't hardcode to true, can we define a top-level variable and pass it down?
Running the rag example notebook will only take around 6-8 minutes in total.