Merge pull request GoogleCloudPlatform#4634 from LAVEEN/a3provision

Enabling Spot VM For A3 Mega/High

Author: LAVEEN

examples/machine-learning/a3-highgpu-8g/README.md

@@ -181,6 +181,34 @@ blueprint matches the name of the user-created reservation.
   a3_maintenance_interval: ""
 ```
 
+### Using Spot VM or DWS Flex
+
+> [!IMPORTANT]
+> Select one of the provisioning models : either spot vm , dws flex or reservation
+
+In order to make use of DWS Flex Start mode with SlurmGCP, you must use the `dws_flex` variable in the `schedmd-slurm-gcp-v6-nodeset` module.
+See the example below:
+
+```yaml
+  vars: 
+    a3_dws_flex_enabled: true             # enabling dws flex by setting the variable to true
+    # the rest of the variables
+```
+
+To learn more about DWS Flex-Start, visit https://github.com/GoogleCloudPlatform/cluster-toolkit/blob/main/docs/slurm-dws-flex.md
+
+Similarly ,to make use of Spot VMs,
+In order to make use of Spot VMs with Slurm, you must use the `enable_spot_vm` variable in the `schedmd-slurm-gcp-v6-nodeset` module.
+See the example below:
+
+```yaml
+  vars: 
+    a3_enable_spot_vm: true             # enabling spot vm by setting the variable to true
+    # the rest of the variables
+```
+
+To learn more about Spot VM  visit: https://cloud.google.com/compute/docs/instances/spot
+
 ### Set cluster size
 
 At approximately line 37 of `ml-slurm-a3-2-cluster.yaml`, set the static cluster

examples/machine-learning/a3-highgpu-8g/ml-slurm-a3-2-cluster.yaml

@@ -39,7 +39,12 @@ vars:
   a3_static_cluster_size: 32
   # a3_reservation_name must be specified; if Google staff have provided you
   # with a reservation name, use it. Otherwise supply user-created reservation.
-  a3_reservation_name: a3-reservation-0
+
+  #Provisioning models (set to true or fill in reservation name, pick only one)
+  a3_reservation_name: ""  # supply reservation name
+  a3_dws_flex_enabled: false
+  a3_enable_spot_vm: false
+
   # a3_maintenance_interval should be empty string by default; if Google staff
   # have created a reservation, they will also provide a3_maintenance_interval
   a3_maintenance_interval: ""
@@ -180,6 +185,9 @@ deployment_groups:
     - a3_startup
     settings:
       reservation_name: $(vars.a3_reservation_name)
+      enable_spot_vm: $(vars.a3_enable_spot_vm)
+      dws_flex:
+        enabled: $(vars.a3_dws_flex_enabled)
       maintenance_interval: $(vars.a3_maintenance_interval)
       node_count_static: $(vars.a3_static_cluster_size)
       node_count_dynamic_max: 0

examples/machine-learning/a3-megagpu-8g/a3mega-slurm-blueprint.yaml

@@ -36,6 +36,11 @@ vars:
   enable_controller_public_ips: true
   localssd_mountpoint: /mnt/localssd
 
+  #Provisioning models (set to true or fill in reservation name, pick only one)
+  a3mega_reservation_name: ""  # supply reservation name
+  a3mega_dws_flex_enabled: false
+  a3mega_enable_spot_vm: false
+
 deployment_groups:
 - group: primary
   modules:
@@ -550,6 +555,9 @@ deployment_groups:
       on_host_maintenance: TERMINATE
       bandwidth_tier: gvnic_enabled
       reservation_name: $(vars.a3mega_reservation_name)
+      enable_spot_vm: $(vars.a3mega_enable_spot_vm)
+      dws_flex:
+        enabled: $(vars.a3mega_dws_flex_enabled)
       maintenance_interval: $(vars.a3mega_maintenance_interval)
       startup_script: $(a3mega_startup.startup_script)
 

examples/machine-learning/a3-megagpu-8g/a3mega-slurm-deployment.yaml

@@ -31,6 +31,9 @@ vars:
   disk_size_gb: 200
   final_image_family: slurm-a3mega
   slurm_cluster_name: a3mega
-  a3mega_reservation_name: "" # supply reservation name
   a3mega_maintenance_interval: ""
   a3mega_cluster_size: 2 # supply cluster size
+  a3mega_reservation_name: "" # supply reservation name
+  # Additional provisioning models (pick only one), can be used to substitute `a3mega_reservation_name`:
+  a3mega_dws_flex_enabled: false # To make use of DWS Flex-Start, for more info visit: https://github.com/GoogleCloudPlatform/cluster-toolkit/blob/main/docs/slurm-dws-flex.md
+  a3mega_enable_spot_vm: false # To make use of Spot VMs, for more info visit: https://cloud.google.com/compute/docs/instances/spot