Console

1. In the Google Cloud console, go to theCreate an instance page.

   Go to Create an instance

2. In theMachine configuration pane, complete the following steps:

   1. In theName field, enter a name for theFlex-start VM.

   2. Specify theRegion andZone where you want to create yourVM. To review the regions and zones where the machine type that youwant to use is available, seeAvailable regions and zones.

   3. Based on the workload that you want to run, specify a machine typeas follows:

      • To specify an accelerator-optimized machine type, do thefollowing:

        1. Click theGPUs tab.

        2. In theGPU type list, select a GPU type, exceptNVIDIA GB200 192GB (A4X) andNVIDIA RTX PRO 600(G4).

        3. In theNumber of GPUs list, select the number of GPUs toattach to your VM.

        4. Optional: If your GPU model supportsNVIDIA RTX Virtual Workstations (vWS) for graphics workloads,and you plan to run graphics-intensive workloads, selectEnable Virtual Workstation (NVIDIA GRID).

      • To specify an H4D machine type, do the following:

        1. Click theCompute optimized tab.

        2. In theSeries column, selectH4D.

3. In the navigation menu, clickAdvanced. In theAdvanced panethat appears, complete the following steps:

   1. In theProvisioning model section, in theVM provisioning model list, selectFlex-start.

   2. In theEnter number of hours field, enter the maximum amount oftime that you want the VM to run. The value must be between0.01(36 seconds) and168 (168 hours, or seven days).

   3. Select theSet a wait time for VM creation checkbox. Then, basedon the zonal requirements for your workload, specify one of thefollowing durations to help increase the chances that your VMcreation request succeeds:

      • If your workload requires you to create the VM in a specificzone, then specify a duration between90 seconds and2 hours. Longer durations give you higher chances ofobtaining resources.

      • If the VM can run in any zone within the region, then specify aduration of0 seconds or clear theSet a wait time for VM creation checkbox. This actionspecifies that Compute Engine only allocates resources if theyare immediately available. If the VM creation request failsbecause resources are unavailable, then retry the request in adifferent zone.

   4. In theOn VM termination field, select whether to stop or deletethe Flex-start VM at the end of its run duration:

      • To delete the VM, selectDelete.

      • To stop the VM, selectStop.

4. To create the Flex-start VM, clickCreate.

gcloud

To create a Flex-start VM, use thegcloud compute instances create commandwith the following flags:

• The--request-valid-for-duration flag

• The--provisioning-model=FLEX_START flag

• The--instance-termination-action flag

• The--max-run-duration flag

• The--maintenance-policy=TERMINATE flag

• The--reservation-affinity=none flag

To create a Flex-start VM, run the following command:

gcloud compute instances createVM_NAME \    --machine-type=MACHINE_TYPE \    --zone=ZONE \    --request-valid-for-duration=VALID_FOR_DURATION \    --provisioning-model=FLEX_START \    --instance-termination-action=TERMINATION_ACTION \    --max-run-duration=RUN_DURATION \    --maintenance-policy=TERMINATE \    --reservation-affinity=none

Replace the following:

• VM_NAME: the name of your new VM.

• MACHINE_TYPE: the machine type to use for theFlex-start VM. If you specify a G2 or N1 machine type,then consider the following:

  • For G2 machine types, you can optionally specify aNVIDIA RTX Virtual Workstations (vWS)to use for graphic-intensive workloads. To do so, include the--accelerator flag in the command as follows:

    --accelerator=count=VWS_ACCELERATOR_COUNT,type=nvidia-l4-vws

    ReplaceVWS_ACCELERATOR_COUNT with the number ofNVIDIA RTX vWS that your workload requires.

  • For N1 machine types, you must specify the number and type of GPUsto attach to your VM. Otherwise, creating the VM fails. To attachGPUs to an N1 VM, include the--accelerator flag in the command asfollows:

    --accelerator=count=NUMBER_OF_ACCELERATORS,type=ACCELERATOR_TYPE

    Replace the following:

    • NUMBER_OF_ACCELERATORS: the number of GPUs toattach to your N1 VM.

    • ACCELERATOR_TYPE: asupported GPU model for N1 VMs.

• ZONE: the zone where you want to create the VM.To verify that your specified machine type is available in the zonewhere you want to create the VM, seeAvailable regions and zones.

• VALID_FOR_DURATION: the maximum time to waitfor provisioning your requested resources. You must format the value asthe number of days, hours, minutes, or seconds followed byd,h,m, ands respectively. For example, specify30m for 30 minutes or1h2m3s for one hour, two minutes, and three seconds. Based on thezonal requirements for your workload, specify one of the followingdurations to help increase your chances that your VM creation requestsucceeds:

  • If your workload requires you to create the VM in a specific zone,then specify a duration between 90 seconds (90s) and two hours(2h). Longer durations give you higher chances of obtainingresources.

  • If the VM can run in any zone within the region, then specify aduration of zero seconds (0s). This value specifies thatCompute Engine only allocates resources if they areimmediately available. If the creation request fails becauseresources are unavailable, then retry the request in a differentzone.

• TERMINATION_ACTION: whether to stop or deletethe VM at the end of its run duration. Specify one of the followingvalues:

  • To stop the VM:STOP

  • To delete the VM:DELETE

• RUN_DURATION: the maximum time that the VM runsbefore Compute Engine automatically stops or deletes it. Youmust format the value as the number of days, hours, minutes, or secondsfollowed byd,h,m, ands respectively. The value must bebetween 10 minutes and seven days.

REST

To create a Flex-start VM, make aPOST request to theinstances.insert method.In the request body, include the following fields:

• Theparams.requestValidForDuration field.

• Thescheduling.provisioningModel field set toFLEX_START.

• Thescheduling.instanceTerminationAction field.

• Thescheduling.maxRunDuration field.

• Thescheduling.onHostMaintenance field set toTERMINATE.

• ThereservationAffinity.consumeReservationType set toNO_RESERVATION.

To create a Flex-start VM, make aPOST request as follows:

POST https://compute.googleapis.com/compute/v1/projects/PROJECT_ID/zones/ZONE/instances{  "name": "VM_NAME",  "machineType": "zones/ZONE/machineTypes/MACHINE_TYPE",  "disks": [    {      "initializeParams": {        "sourceImage": "projects/IMAGE_PROJECT/global/images/IMAGE"      },      "boot": true    }  ],  "networkInterfaces": [    {      "network": "global/networks/default"    }  ],  "params": {    "requestValidForDuration": {      "seconds":VALID_FOR_DURATION    }  },  "scheduling": {    "provisioningModel": "FLEX_START",    "instanceTerminationAction": "TERMINATION_ACTION",    "maxRunDuration": {      "seconds":RUN_DURATION    },    "onHostMaintenance": "TERMINATE"  },  "reservationAffinity": {    "consumeReservationType": "NO_RESERVATION"  }}

Replace the following:

• PROJECT_ID: the ID of the project in which tocreate the VM.

• ZONE: the zone where you want to create the VM.To verify that a machine type is available in the zone where you want tocreate the VM, seeAvailable regions and zones.

• VM_NAME: the name of your new VM.

• MACHINE_TYPE: the machine type to use for theFlex-start VM. If you specify a G2 or N1 machine type,then consider the following:

  • For G2 machine types, you can optionally specify aNVIDIA RTX Virtual Workstations (vWS)to use for graphic-intensive workloads. To do so, include theguestAccelerators field in the request body as follows:

    "guestAccelerators": [  {    "acceleratorCount":VWS_ACCELERATOR_COUNT,    "acceleratorType": "projects/PROJECT_ID/zones/ZONE/acceleratorTypes/nvidia-l4-vws"  }]

    ReplaceVWS_ACCELERATOR_COUNT with the number ofNVIDIA RTX vWS that your workload requires.

  • For N1 machine types, you must specify the number and type of GPUsto attach to your VM. Otherwise, creating the VM fails. To attachGPUs to an N1 VM, include theguestAccelerators field in therequest body as follows:

    "guestAccelerators": [  {    "acceleratorCount":ACCELERATOR_COUNT,    "acceleratorType": "projects/PROJECT_ID/zones/ZONE/acceleratorTypes/ACCELERATOR_TYPE"  }]

    Replace the following:

    • NUMBER_OF_ACCELERATORS: the number of GPUs toattach to your N1 VM.

    • ACCELERATOR_TYPE: asupported GPU model for N1 VMs.

• IMAGE_PROJECT: the image project that containsthe image—for example,debian-cloud. For more information aboutthe supported image projects, seePublic images.

• IMAGE: specify one of the following:

  • A specific version of the OS image—for example,debian-12-bookworm-v20240617.

  • Animage family, which must beformatted asfamily/IMAGE_FAMILY. This valuespecifies to use the most recent, non-deprecated OS image. Forexample, if you specifyfamily/debian-12, the latest version inthe Debian 12 image family is used. For more information about usingimage families, seeImage families best practices.

• VALID_FOR_DURATION: the maximum time in secondsto wait for the VM to be provisioned. Based on the zonal requirementsfor your workload, specify one of the following durations to helpincrease your chances that your VM creation request succeeds:

  • If your workload requires you to create the VM in a specific zone,then specify a duration between 90 seconds (90) and two hours(7200). Longer durations give you higher chances of obtainingresources.

  • If the VM can run in any zone within the region, then specify aduration of zero seconds (0). This value specifies thatCompute Engine only allocates resources if they areimmediately available. If the creation request fails becauseresources aren't available, then retry the request in a differentzone.

• TERMINATION_ACTION: whether to stop or deletethe VM at the end of its run duration. Specify one of the followingvalues:

  • To stop the VM:STOP

  • To delete the VM:DELETE

• RUN_DURATION: the maximum time, in seconds,that the VM runs before Compute Engine automatically stops ordeletes it. The value must be between600 (600 seconds, or 10 minutes)and604800 (604,800 seconds, or seven days).