For workloads that require temporary storage with high performance and low
latency, consider using local solid-state drive (Local SSD) disks when you
create your compute instance. Local SSD disks are always-encrypted temporary
solid-state storage for Compute Engine. To learn about the other disks
available in Compute Engine, see
[Choose a disk type](https://docs.cloud.google.com/compute/docs/disks.md.txt).

Local SSD disks are ideal when you need storage for any of the following use
cases:

- Caches or storage for transient data
- Scratch processing space for high performance computing or data analytics
- Temporary data storage like for the `tempdb` system database for Microsoft SQL Server

Local SSD disks offer superior I/O operations per second (IOPS), and very low
latency compared to the persistent storage provided by
[Google Cloud Hyperdisk](https://docs.cloud.google.com/compute/docs/disks/hyperdisks.md.txt) and
[Persistent Disk](https://docs.cloud.google.com/compute/docs/disks/persistent-disks.md.txt). This low latency is
because Local SSD disks are physically attached to the server that hosts your
instance. For this same reason, Local SSD disks can
**provide only temporary storage**.

> [!WARNING]
> **Warning:** Local SSD storage is **not** automatically replicated and **all data on the Local SSD may be lost** if the instance stops or is terminated for any reason. See [Local SSD data persistence](https://docs.cloud.google.com/compute/docs/disks/local-ssd.md.txt#data_persistence) for details.

Because Local SSD is suitable only for temporary storage, you must store data
that isn't temporary or ephemeral in nature on a Hyperdisk
or Persistent Disk volume.

To use Local SSD disks with a compute instance,
[add Local SSD disks when you create the instance](https://docs.cloud.google.com/compute/docs/disks/add-local-ssd.md.txt).
You can't add Local SSD disks to an instance after creating the instance.

## Limitations

Local SSD has the following limitations:

- You can't use Local SSD disks with virtual machine (VM) instances that use
  [shared-core](https://docs.cloud.google.com/compute/docs/machine-resource.md.txt#shared-core-types) machine
  types.

- Not all machine series and machine types support Local SSD. For a full list of
  machine series that support Local SSD, see
  [Machine series support](https://docs.cloud.google.com/compute/docs/disks/local-ssd.md.txt#machine-series-lssd).

- You can't use customer-supplied encryption keys or customer-managed encryption
  keys with Local SSD disks. Compute Engine automatically encrypts
  your data when it's written to Local SSD storage.

- You can't back up Local SSD disks with snapshots, clones, machine images, or
  images. Store important data on Hyperdisk or Persistent Disk
  volumes.

## Types of Local SSD disks

Local SSD disks come in two types:

- **Titanium SSD** : Titanium SSD is a custom-designed Local SSD disk
  that uses [Titanium I/O offload processing](https://cloud.google.com/titanium)
  and offers enhanced SSD security, performance, and management.
  Titanium offers higher storage IOPS, throughput, and lower latency
  than the previous generation of Local SSD. The following machine series offer
  Local SSD storage using Titanium SSD:

  - General-purpose: [C4](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c4_series), [C4A](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c4a_series), and [C4D](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c4d_series) machine series
  - Storage-optimized: [Z4D](https://docs.cloud.google.com/compute/docs/storage-optimized-machines.md.txt#z4d_series), [Z4M](https://docs.cloud.google.com/compute/docs/storage-optimized-machines.md.txt#z4m_series), and [Z3](https://docs.cloud.google.com/compute/docs/storage-optimized-machines.md.txt#z3_series) machine series
  - Compute-optimized: [H4D](https://docs.cloud.google.com/compute/docs/compute-optimized-machines.md.txt#h4d_series) machine series
  - Network-optimized: [C4N](https://docs.cloud.google.com/compute/docs/network-optimized-machines.md.txt#c4n_series)
  - Accelerator-optimized: [A5X](https://docs.cloud.google.com/compute/docs/accelerator-optimized-machines.md.txt#a5x-instances), [A4X Max and A4X](https://docs.cloud.google.com/compute/docs/accelerator-optimized-machines.md.txt#a4x-vms), and [G4](https://docs.cloud.google.com/compute/docs/accelerator-optimized-machines.md.txt#g4-series) machine series

  Titanium SSD disks are directly attached to the compute instances
  inside their host server.
- **Local SSD**: Local SSD is the original local SSD feature for Google Cloud.
  Each Local SSD disk attached to an instance provides 375 GiB of capacity.
  These disks provide higher performance than Hyperdisk or
  Persistent Disk. You can use either the NVMe or SCSI interface to mount Local
  SSD disks.

  Local SSD disks are directly attached to the instances inside their host
  server.

Unless Titanium SSD is specifically mentioned, the term
"Local SSD" applies to both Local SSD and Titanium SSD when describing
features of local SSD disks.

## Performance

Local SSD performance depends on several factors, including the number of
attached Local SSD disks, the selected disk interface
([NVMe](http://wikipedia.org/wiki/NVM_Express) or
[SCSI](http://wikipedia.org/wiki/SCSI)), and the instance's
machine type. The available performance increases as you attach more Local SSD
disks to your instance.

### Local SSD performance by number of attached disks

The following tables list the maximum IOPS and throughput for NVMe- and
SCSI-attached Local SSD disks. The metrics are listed by the total capacity of
Local SSD disks attached to the instance.

#### Titanium SSD performance

The following table lists the performance limits for Titanium SSD disks on
[C4](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c4_series),
[C4A](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c4a_series),
[C4D](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c4d_series),
[C4N](https://docs.cloud.google.com/compute/docs/network-optimized-machines.md.txt#c4n_series),
[G4](https://docs.cloud.google.com/compute/docs/accelerator-optimized-machines.md.txt#g4-series),
[H4D](https://docs.cloud.google.com/compute/docs/compute-optimized-machines.md.txt#h4d_series),
[Z3](https://docs.cloud.google.com/compute/docs/storage-optimized-machines.md.txt#z3_series),
[Z4D](https://docs.cloud.google.com/compute/docs/storage-optimized-machines.md.txt#z4d_series), and
[Z4M](https://docs.cloud.google.com/compute/docs/storage-optimized-machines.md.txt#z4m_series) instances.

| Machine type | # of attached Titanium SSD disks | Total storage space (GiB) | IOPS || Throughput (MiB/s) ||
|   ||| Read | Write | Read | Write |
|---|---|---|---|---|---|---|
| C4 (with 375 GiB disks) |||||||
| `c4-standard-4-lssd` `c4-highmem-4-lssd` | 1 | 375 | 150,000 | 75,000 | 625 | 330 |
| `c4-standard-8-lssd` `c4-highmem-8-lssd` | 1 | 375 | 150,000 | 75,000 | 625 | 330 |
| `c4-standard-16-lssd` `c4-highmem-16-lssd` | 2 | 750 | 300,000 | 150,000 | 1,250 | 660 |
| `c4-standard-24-lssd` `c4-highmem-24-lssd` | 4 | 1,500 | 600,000 | 300,000 | 2,500 | 1,320 |
| `c4-standard-32-lssd` `c4-highmem-32-lssd` | 5 | 1,875 | 750,000 | 375,000 | 3,125 | 1,650 |
| `c4-standard-48-lssd` `c4-highmem-48-lssd` | 8 | 3,000 | 1,200,000 | 600,000 | 5,000 | 2,640 |
| `c4-standard-96-lssd` `c4-highmem-96-lssd` | 16 | 6,000 | 2,400,000 | 1,200,000 | 10,000 | 5,280 |
| `c4-standard-144-lssd` `c4-highmem-144-lssd` | 24 | 9,000 | 3,600,000 | 1,800,000 | 15,000 | 7,920 |
| `c4-standard-192-lssd` `c4-highmem-192-lssd` | 32 | 12,000 | 4,800,000 | 2,400,000 | 20,000 | 10,560 |
| `c4-standard-288-lssd` `c4-highmem-288-lssd` | 48 | 18,000 | 7,200,000 | 3,600,000 | 30,000 | 15,840 |
| C4 bare metal (with 3,000 GiB disks) |||||||
| `c4-standard-288-lssd-metal` `c4-highmem-288-lssd-metal` | 6 | 18,000 | 7,200,000 | 3,600,000 | 30,000 | 15,840 |
| C4A (with 375 GiB disks) |||||||
| `c4a-standard-4-lssd` `c4a-highmem-4-lssd` | 1 | 375 | 150,000 | 75,000 | 650 | 330 |
| `c4a-standard-8-lssd` `c4a-highmem-8-lssd` | 2 | 750 | 300,000 | 150,000 | 1,300 | 660 |
| `c4a-standard-16-lssd` `c4a-highmem-16-lssd` | 4 | 1,500 | 600,000 | 300,000 | 2,600 | 1,320 |
| `c4a-standard-32-lssd` `c4a-highmem-32-lssd` | 6 | 2,250 | 900,000 | 450,000 | 3,900 | 1,980 |
| `c4a-standard-48-lssd` `c4a-highmem-48-lssd` | 10 | 3,750 | 1,500,000 | 750,000 | 6,500 | 3,300 |
| `c4a-standard-64-lssd` `c4a-highmem-64-lssd` | 14 | 5,250 | 2,100,000 | 1,050,000 | 9,100 | 4,620 |
| `c4a-standard-72-lssd` `c4a-highmem-72-lssd` | 16 | 6,000 | 2,400,000 | 1,200,000 | 10,400 | 5,280 |
| C4D (with 375 GiB disks) |||||||
| `c4d-standard-8-lssd` `c4d-highmem-8-lssd` `c4d-standard-16-lssd` `c4d-highmem-16-lssd` | 1 | 375 | 150,000 | 75,000 | 625 | 330 |
| `c4d-standard-32-lssd` `c4d-highmem-32-lssd` | 2 | 750 | 300,000 | 150,000 | 1,250 | 660 |
| `c4d-standard-48-lssd` `c4d-highmem-48-lssd` | 4 | 1,500 | 600,000 | 300,000 | 2,500 | 1,320 |
| `c4d-standard-64-lssd` `c4d-highmem-64-lssd` | 6 | 2,250 | 900,000 | 450,000 | 3,750 | 1,980 |
| `c4d-standard-96-lssd` `c4d-highmem-96-lssd` | 8 | 3,000 | 1,200,000 | 600,000 | 5,000 | 2,640 |
| `c4d-standard-192-lssd` `c4d-highmem-192-lssd` | 16 | 6,000 | 2,400,000 | 1,200,000 | 10,000 | 5,280 |
| `c4d-standard-384-lssd` `c4d-highmem-384-lssd` | 32 | 12,000 | 4,800,000 | 2,400,000 | 20,000 | 10,560 |
| C4N (with 375 GiB disks) |||||||
| `c4n-highmem-4-lssd` `c4n-highmem-8-lssd` | 1 | 375 | 150,000 | 75,000 | 625 | 330 |
| `c4n-highmem-16-lssd` | 2 | 750 | 300,000 | 150,000 | 1,250 | 660 |
| `c4n-highmem-24-lssd` | 4 | 1,500 | 600,000 | 300,000 | 2,500 | 1,320 |
| `c4n-highmem-48-lssd` | 8 | 3,000 | 1,200,000 | 600,000 | 5,000 | 2,640 |
| `c4n-highmem-96-lssd` | 16 | 6,000 | 2,400,000 | 1,200,000 | 10,000 | 5,280 |
| `c4n-highmem-192-lssd` | 32 | 12,000 | 4,800,000 | 2,400,000 | 20,000 | 10,560 |
| G4 (with 375 GiB disks) |||||||
| `g4-standard-48` | 4 | 1,500 | 600,000 | 300,000 | 2,500 | 1,320 |
| `g4-standard-96` | 8 | 3,000 | 1,200,000 | 600,000 | 5,000 | 2,640 |
| `g4-standard-192` | 16 | 6,000 | 2,400,000 | 1,200,000 | 10,000 | 5,280 |
| `g4-standard-384` | 32 | 12,000 | 4,800,000 | 2,400,000 | 20,000 | 10,560 |
| H4D (with 375 GiB disks) |||||||
| `h4d-highmem-192-lssd` | 10 | 3,750 | 1,500,000 | 750,000 | 6,250 | 3,300 |
| `h4d-megamem-192-lssd` | 32 | 12,000 | 4,800,000 | 2,400,000 | 20,000 | 10,560 |
| Z3 (with 3,000 GiB disks) |||||||
| `z3-highmem-8-highlssd` `z3-highmem-14-standardlssd` | 1 | 3,000 | 750,000 | 500,000 | 3,000 | 2,500 |
| `z3-highmem-16-highlssd` `z3-highmem-22-standardlssd` | 2 | 6,000 | 1,500,000 | 1,000,000 | 6,000 | 5,000 |
| `z3-highmem-22-highlssd` `z3-highmem-44-standardlssd` | 3 | 9,000 | 2,250,000 | 1,500,000 | 9,000 | 7,500 |
| `z3-highmem-32-highlssd` | 4 | 12,000 | 3,000,000 | 2,000,000 | 12,000 | 10,000 |
| `z3-highmem-44-highlssd` `z3-highmem-88-standard` | 6 | 18,000 | 4,500,000 | 3,000,000 | 18,000 | 15,000 |
| `z3-highmem-88-highlssd` `z3-highmem-176-standardlssd` | 12 | 36,000 | 9,000,000 | 6,000,000 | 36,000 | 30,000 |
| Z3 bare metal (with 6,000 GiB disks) |||||||
| `z3-highmem-192-highlssd-metal` | 12 | 72,000 | 9,000,000 | 6,000,000 | 36,000 | 30,000 |
| Z4D (with 3,500 GiB disks) |||||||
| `z4d-highmem-16-standardlssd` `z4d-highmem-8-highlssd` | 1 | 3,500 | 650,000 | 512,000 | 3,150 | 2,400 |
| `z4d-highmem-32-standardlssd` `z4d-highmem-16-highlssd` | 2 | 7,000 | 1,300,000 | 1,024,000 | 6,300 | 4,800 |
| `z4d-highmem-48-standardlssd` | 3 | 10,500 | 1,950,000 | 1,536,000 | 9,450 | 7,200 |
| `z4d-highmem-64-standardlssd` `z4d-highmem-32-highlssd` | 4 | 14,000 | 2,600,000 | 2,048,000 | 12,600 | 9,600 |
| `z4d-highmem-96-standardlssd` `z4d-highmem-48-highlssd` | 6 | 21,000 | 3,900,000 | 3,072,000 | 18,900 | 14,400 |
| `z4d-highmem-64-highlssd` | 8 | 28,000 | 5,200,000 | 4,096,000 | 25,200 | 19,200 |
| `z4d-highmem-192-standardlssd` `z4d-highmem-96-highlssd` | 12 | 42,000 | 7,800,000 | 6,144,000 | 37,800 | 28,800 |
| `z4d-highmem-384-standardlssd` `z4d-highmem-192-highlssd` | 24 | 84,000 | 15,600,000 | 12,288,000 | 75,600 | 57,600 |
| Z4D bare metal (with 7,000 GiB disks) |||||||
| `z4d-highmem-384-standardlssd-metal` | 12 | 84,000 | 15,600,000 | 12,288,000 | 75,600 | 57,600 |
| Z4M (with 14,000 GiB disks) |||||||
| `z4m-highmem-16` | 1 | 14,000 | 1,300,000 | 1,024,000 | 6,300 | 4,800 |
| `z4m-highmem-32` | 2 | 28,000 | 2,600,000 | 2,048,000 | 12,600 | 9,600 |
| `z4m-highmem-48` | 3 | 42,000 | 3,900,000 | 3,072,000 | 18,900 | 14,400 |
| `z4m-highmem-192` | 12 | 168,000 | 15,600,000 | 12,288,000 | 75,600 | 57,600 |
| `z4m-highmem-192-metal` | 12 | 168,000 | 15,600,000 | 12,288,000 | 75,600 | 57,600 |

#### NVMe Local SSD performance

The following table lists the performance limits for Local SSD disks that are
attached to instances using NVMe.

| # of attached Local SSD disks | Total storage space (GiB) | Capacity per disk (GiB) | IOPS || Throughput (MiB/s) ||
|   ||| Read | Write | Read | Write |
|---|---|---|---|---|---|---|
| 1 | 375 | 375 | 170,000 | 90,000 | 660 | 350 |
| 2 | 750 | 375 | 340,000 | 180,000 | 1,320 | 700 |
| 3 | 1,125 | 375 | 510,000 | 270,000 | 1,980 | 1,050 |
| 4 | 1,500 | 375 | 680,000 | 360,000 | 2,650 | 1,400 |
| 5 | 1,875 | 375 | 680,000 | 360,000 | 2,650 | 1,400 |
| 6 | 2,250 | 375 | 680,000 | 360,000 | 2,650 | 1,400 |
| 7 | 2,625 | 375 | 680,000 | 360,000 | 2,650 | 1,400 |
| 8 | 3,000 | 375 | 680,000 | 360,000 | 2,650 | 1,400 |
| 16 | 6,000 | 375 | 1,600,000 | 800,000 | 6,240 | 3,120 |
| 24 | 9,000 | 375 | 2,400,000 | 1,200,000 | 9,360 | 4,680 |
| 32 | 12,000 | 375 | 3,200,000 | 1,600,000 | 12,480 | 6,240 |

#### SCSI Local SSD performance

The following table lists the performance limits for Local SSD disks that are
attached to instances using SCSI.

| # of combined Local SSD disks | Storage space (GiB) | IOPS || Throughput (MiB/s) ||
|   || Read | Write | Read | Write |
|---|---|---|---|---|---|
| 1 | 375 | 100,000 | 70,000 | 390 | 270 |
| 2 | 750 | 200,000 | 140,000 | 780 | 550 |
| 3 | 1,125 | 300,000 | 210,000 | 1,170 | 820 |
| 4 | 1,500 | 400,000 | 280,000 | 1,560 | 1,090 |
| 5 | 1,875 | 400,000 | 280,000 | 1,560 | 1,090 |
| 6 | 2,250 | 400,000 | 280,000 | 1,560 | 1,090 |
| 7 | 2,625 | 400,000 | 280,000 | 1,560 | 1,090 |
| 8 | 3,000 | 400,000 | 280,000 | 1,560 | 1,090 |
| 16 | 6,000 | 900,000 | 800,000 | 6,240 | 3,120 |
| 24 | 9,000 | 900,000 | 800,000 | 9,360 | 4,680 |

### Configure your instance to maximize performance

To reach the stated performance levels, you must configure your compute instance
as follows:

- Attach the Local SSD disks with the NVMe interface. Disks attached using the
  SCSI interface have lower performance.

- The following machine types also require a minimum number of vCPUs to reach
  these maximums:

  - [N2](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#n2_series), [N2D](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#n2d_machines), or [A2](https://docs.cloud.google.com/compute/docs/accelerator-optimized-machines.md.txt#a2-vms) machine types require at least 24 vCPUs.
  - [N1](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#n1_machines) machine types require at least 32 vCPUs.
- If your instance uses a custom Linux image, then the image must use version
  4.14.68 or later of the Linux kernel. If you use the public images provided by
  Compute Engine, then you don't have to take any further action.

For additional instance and disk configuration settings that can improve Local
SSD performance, see
[Optimizing Local SSD performance](https://docs.cloud.google.com/compute/docs/disks/optimizing-local-ssd-performance.md.txt).

For more information about selecting a disk interface,
see [Choose a disk interface](https://docs.cloud.google.com/compute/docs/disks/local-ssd.md.txt#choose_an_interface).

## Local SSD data persistence

Compute Engine preserves the data on Local SSD disks only in certain
scenarios.

### Scenarios where Compute Engine persists Local SSD data

Compute Engine preserves the data on Local SSD disks in the
following events:

- You reboot the guest operating system (OS).
- You configure your instance for [live migration](https://docs.cloud.google.com/compute/docs/instances/live-migration-process.md.txt) and the instance goes through a host maintenance event.
- You opt to [preserve the Local SSD data](https://docs.cloud.google.com/compute/docs/disks/local-ssd.md.txt#stop_instance) when you stop or suspend the instance ([Preview](https://cloud.google.com/products/#product-launch-stages)).
- A compute instance with attached Local SSD disks that supports only termination and automatic restart undergoes a maintenance event. The instance restarts in place, preserving the Local SSD data, instead of migrating to a new host.

### Scenarios where Compute Engine might not persist Local SSD data

Data on Local SSD disks might be lost if a
[host error](https://docs.cloud.google.com/compute/docs/faq.md.txt#hosterror) occurs on the instance and
Compute Engine can't recover the data on the Local SSD disks attached to
the instance within a specified time.

You can control how much time, if any, is spent attempting to recover
the data with [Local SSD recovery timeout](https://docs.cloud.google.com/compute/docs/instances/host-maintenance-overview.md.txt#local-ssd-timeout).
Recovery is attempted only if your instance is configured to automatically
restart (`automaticRestart`) after the repair operation completes.

If Compute Engine *can* recover the data on the Local SSD disks, the
instance restarts with all disks attached, and the recovered disk states will be
consistent with the disk states prior to the failure.

> [!WARNING]
> **Warning:** Local SSD disks attached to C4, C4A, C4D, C4N, and H4D instances might not recover all writes that completed before a power loss event. For more information, see [Local SSD data loss on C4, C4A, C4D, C4N, and H4D instances](https://docs.cloud.google.com/compute/docs/troubleshooting/known-issues.md.txt#lssd-power-loss).

If Compute Engine *can't* recover the data on the Local SSD disks before
the timeout expires, the instance restarts with blank, unformatted Local SSD
disks attached, and the original data is unrecoverable.

### Scenarios where Compute Engine doesn't persist Local SSD data

Data on Local SSD disks doesn't persist through the following events:

- You shut down the guest OS and force the instance to stop.
- You create a [Spot VM](https://docs.cloud.google.com/compute/docs/instances/spot.md.txt) or [preemptible VM](https://docs.cloud.google.com/compute/docs/instances/preemptible.md.txt) and the VM goes through the preemption process.
- You configure the instance to [stop on host maintenance events](https://docs.cloud.google.com/compute/docs/instances/host-maintenance-overview.md.txt#terminate_and_optionally_restart) instead of using live migration, and the instance goes through a host maintenance event.
- You [limit the run time of an instance](https://docs.cloud.google.com/compute/docs/instances/limit-vm-runtime.md.txt), select `STOP` as the termination action, and the instance reaches its time limit.
- You misconfigure a Local SSD disk and it becomes unreachable.
- You disable project billing, causing the instance to stop.
- If `automaticRestart` isn't configured on your instance.

If Compute Engine can't recover an instance's Local SSD data,
the instance restarts with blank Local SSD disks attached, and the
original data is unrecoverable.

## Machine series support

You can use Local SSD disks with the following machine series.

> [!NOTE]
> **Note:** You can use Local SSD disks with a [bare metal instance](https://docs.cloud.google.com/compute/docs/instances/bare-metal-instances.md.txt) only if the machine type name contains both `metal` and `lssd`, or if the instance is an A5X or A4X Max bare metal instance, which automatically includes Local SSD disks.


A2 A3 (H100) A3 (H200) A4 A4X A4X Max A5X C2 C2D C3 C3D C4 C4A C4D C4N E2 G2 G4 H3 H4D M1 M2 M3 M4 M4N N1 N1+GPU N2 N2D N4 N4A N4D T2A T2D TPU v2 TPU v3 TPU v4 TPU v5e TPU v5p TPU v6e TPU7x X4 X5 Z3 Z4D Z4M
Select a machine series to display its support for Local SSD.

| Machine series | Local SSD support |
|---|---|
| [A2](https://docs.cloud.google.com/compute/docs/accelerator-optimized-machines.md.txt#a2-vms) | Yes |
| [A3 (H100)](https://docs.cloud.google.com/compute/docs/gpus.md.txt#h100-gpus) | Yes |
| [A3 (H200)](https://docs.cloud.google.com/compute/docs/gpus.md.txt#h200-gpus) | Yes |
| [A4](https://docs.cloud.google.com/compute/docs/accelerator-optimized-machines.md.txt#a4-vms) | Yes |
| [A4X](https://docs.cloud.google.com/compute/docs/accelerator-optimized-machines.md.txt#a4x-vms) | Yes |
| [A4X Max](https://docs.cloud.google.com/compute/docs/accelerator-optimized-machines.md.txt#a4x-vms) | Yes |
| [A5X](https://docs.cloud.google.com/compute/docs/accelerator-optimized-machines.md.txt#a5x-instances) | Yes |
| [C2](https://docs.cloud.google.com/compute/docs/compute-optimized-machines.md.txt#c2_machine_types) | Yes |
| [C2D](https://docs.cloud.google.com/compute/docs/compute-optimized-machines.md.txt#c2d_machine_types) | Yes |
| [C3](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c3_series) | Yes |
| [C3D](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c3d_machine_types) | Yes |
| [C4](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c4_series) | Yes |
| [C4A](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c4a_series) | Yes |
| [C4D](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c4d_series) | Yes |
| [C4N](https://docs.cloud.google.com/compute/docs/network-optimized-machines.md.txt#c4n_series) | Yes |
| [E2](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#e2_machine_types) | **---** |
| [G2](https://docs.cloud.google.com/compute/docs/accelerator-optimized-machines.md.txt#g2-vms) | Yes |
| [G4](https://docs.cloud.google.com/compute/docs/accelerator-optimized-machines.md.txt#g4-series) | Yes |
| [H3](https://docs.cloud.google.com/compute/docs/compute-optimized-machines.md.txt#h3_series) | **---** |
| [H4D](https://docs.cloud.google.com/compute/docs/compute-optimized-machines.md.txt#h4d_series) | Yes |
| [M1](https://docs.cloud.google.com/compute/docs/memory-optimized-machines.md.txt#m1_machine_types) | Yes |
| [M2](https://docs.cloud.google.com/compute/docs/memory-optimized-machines.md.txt#m2_machine_types) | **---** |
| [M3](https://docs.cloud.google.com/compute/docs/memory-optimized-machines.md.txt#m3_machine_types) | Yes |
| [M4](https://docs.cloud.google.com/compute/docs/memory-optimized-machines.md.txt#m4_machine_types) | **---** |
| [M4N](https://docs.cloud.google.com/compute/docs/network-optimized-machines.md.txt#m4n_series) | **---** |
| [N1](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#n1_machines) | Yes |
| [N1+GPU](https://docs.cloud.google.com/compute/docs/gpus.md.txt) | Yes |
| [N2](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#n2_series) | Yes |
| [N2D](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#n2d_machines) | Yes |
| [N4](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#n4_series) | **---** |
| [N4A](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#n4a_series) | **---** |
| [N4D](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#n4d_series) | **---** |
| [T2A](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#t2a_machines) | **---** |
| [T2D](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#t2d_machines) | **---** |
| [TPU v2](https://docs.cloud.google.com/tpu/docs/v2.md.txt) | **---** |
| [TPU v3](https://docs.cloud.google.com/tpu/docs/v3.md.txt) | **---** |
| [TPU v4](https://docs.cloud.google.com/tpu/docs/v4.md.txt) | **---** |
| [TPU v5e](https://docs.cloud.google.com/tpu/docs/v5e.md.txt) | **---** |
| [TPU v5p](https://docs.cloud.google.com/tpu/docs/v5p.md.txt) | **---** |
| [TPU v6e](https://docs.cloud.google.com/tpu/docs/v6e.md.txt) | **---** |
| [TPU7x](https://docs.cloud.google.com/tpu/docs/tpu7x.md.txt) | **---** |
| [X4](https://docs.cloud.google.com/compute/docs/memory-optimized-machines.md.txt#x4_series) | **---** |
| [X5](https://docs.cloud.google.com/compute/docs/memory-optimized-machines.md.txt#x5_series) | **---** |
| [Z3](https://docs.cloud.google.com/compute/docs/storage-optimized-machines.md.txt#z3_series) | Yes |
| [Z4D](https://docs.cloud.google.com/compute/docs/storage-optimized-machines.md.txt#z4d_series) | Yes |
| [Z4M](https://docs.cloud.google.com/compute/docs/storage-optimized-machines.md.txt#z4m_series) | Yes |

However, there are constraints around how many Local SSD disks you can
attach based on each machine type. For more information, see
[Choose a valid number of Local SSD disks](https://docs.cloud.google.com/compute/docs/disks/local-ssd.md.txt#choose_number_local_ssds).

## Local SSD encryption

Compute Engine automatically encrypts your data when it is written to
Local SSD disks. You can't use
[customer-supplied encryption keys](https://docs.cloud.google.com/compute/docs/disks/customer-supplied-encryption.md.txt)
with Local SSD disks.

## Local SSD data backup

Since you can't back up Local SSD data with disk images, standard snapshots, or
disk clones, Google recommends that you always store valuable data on a
[durable storage option](https://docs.cloud.google.com/compute/docs/disks.md.txt).

If you need to preserve the data on a Local SSD disk, attach a Persistent Disk or
Google Cloud Hyperdisk to the instance. After you mount the Persistent Disk or
Hyperdisk copy the data from the Local SSD disk to the newly
attached disk.

## Choose a disk interface

To achieve the highest Local SSD performance, you must attach your disks to the
instance with the NVMe interface. Performance is lower if you use the SCSI
interface.

The disk interface you choose also depends on the machine type and OS that your
instance uses. Some of the available machine types in Compute Engine
allow you to choose between NVMe and SCSI interfaces, while others support
either only NVMe or only SCSI. Similarly, some of the public OS images provided
by Compute Engine might support both NVMe and SCSI, or only one of the
two.

### Disk interface support by machine type and OS image

The following pages provide more information about available machine types and
supported public images, as well as performance details.

- **Supported interfaces by machine types** : See
  [Machine series comparison](https://docs.cloud.google.com/compute/docs/machine-resource.md.txt#machine_type_comparison).
  In the **Choose instance properties to compare** list, select
  **Disk interface type**.

- **OS image** : For a list of which public OS images provided by
  Compute Engine support SCSI or NVMe, see the
  [Interfaces](https://docs.cloud.google.com/compute/docs/images/os-details.md.txt#interfaces) tab for each table
  in the operating system details documentation.

### Considerations for NVMe for custom images

If your instance uses a custom Linux image, you must use version 4.14.68 or
later of the Linux kernel for optimal NVMe performance.

If your instance uses Windows, and your instance was created before May 2022,
you must replace your NVMe driver. For more information, see
[Replace the NVMe driver on VMs created before May 2022](https://docs.cloud.google.com/compute/docs/troubleshooting/known-issues.md.txt#replace-community-nvme-driver).

### Considerations for SCSI for custom images

If you have an existing setup that requires using a SCSI interface, consider
using multi-queue SCSI to achieve better performance over the standard SCSI
interface.

If you are using a custom image that you imported, see
[Enable multi-queue SCSI](https://docs.cloud.google.com/compute/docs/disks/optimizing-local-ssd-performance.md.txt#multiqueue).

## Choose a valid number of Local SSD disks

Some machine types always include a fixed number of Local SSD disks by default,
while others allow you to add specific numbers of disks. You can only add Local
SSD disks when you create the instance. You can't add Local SSD disks to an
instance after you create it.

The size of each individual Local SSD disk varies by machine series, as follows:

| Machine series or type | Local or Titanium SSD disk size |
|---|---|
| A5X bare metal instances | 12,000 GiB |
| A4X VM instances, A4X Max bare metal instances | 3,000 GiB |
| Z3 VM instances | 3,000 GiB |
| Z3 bare metal instances | 6,000 GiB |
| Z4D VM instances | 3,500 GiB |
| Z4D bare metal instances | 7,000 GiB |
| Z4M VM instances | 14,000 GiB |
| All others | 375 GiB |

### Machine types that automatically attach Local SSD disks

The following table lists the machine types that come with Local SSD disks by
default. The table also shows the number of these disks that are attached when
you create the instance.

| Machine type | Number of Local SSD disks automatically attached per instance |
|---|---|---|
| A5X ||
| `a5x-highgpu-4g-metal` | 4 |
| A4X Max ||
| `a4x-maxgpu-4g-metal` | 4 |
| A4X ||
| `a4x-highgpu-4g` | 4 |
| A4 ||
| `a4-highgpu-8g` | 32 |
| A3 Ultra ||
| `a3-ultragpu-8g` | 32 |
| A3 Mega ||
| `a3-megagpu-8g` | 16 |
| A3 High ||
| `a3-highgpu-1g` | 2 |
| `a3-highgpu-2g` | 4 |
| `a3-highgpu-4g` | 8 |
| `a3-highgpu-8g` | 16 |
| A3 Edge ||
| `a3-edgegpu-8g` | 16 |
| A2 Ultra ||
| `a2-ultragpu-1g` | 1 |
| `a2-ultragpu-2g` | 2 |
| `a2-ultragpu-4g` | 4 |
| `a2-ultragpu-8g` | 8 |
| C4 using Titanium SSD ||
| *Only the [`-lssd` variants of the C4 machine types](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c4_series) support local Titanium SSD.* ||
| `c4-standard-4-lssd` `c4-highmem-4-lssd` | 1 |
| `c4-standard-8-lssd` `c4-highmem-8-lssd` | 1 |
| `c4-standard-16-lssd` `c4-highmem-16-lssd` | 2 |
| `c4-standard-24-lssd` `c4-highmem-24-lssd` | 4 |
| `c4-standard-32-lssd` `c4-highmem-32-lssd` | 5 |
| `c4-standard-48-lssd` `c4-highmem-48-lssd` | 8 |
| `c4-standard-96-lssd` `c4-highmem-96-lssd` | 16 |
| `c4-standard-144-lssd` `c4-highmem-144-lssd` | 24 |
| `c4-standard-192-lssd` `c4-highmem-192-lssd` | 32 |
| `c4-standard-288-lssd` `c4-highmem-288-lssd` | 48 |
| `c4-standard-288-lssd-metal` `c4-highmem-288-lssd-metal` | 6 |
| C4A using Titanium SSD ||
| *Only the [`-lssd` variants of the C4A machine types](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c4a_machine_types) support local Titanium SSD.* ||
| `c4a-standard-4-lssd` `c4a-highmem-4-lssd` | 1 |
| `c4a-standard-8-lssd` `c4a-highmem-8-lssd` | 2 |
| `c4a-standard-16-lssd` `c4a-highmem-16-lssd` | 4 |
| `c4a-standard-32-lssd` `c4a-highmem-32-lssd` | 6 |
| `c4a-standard-48-lssd` `c4a-highmem-48-lssd` | 10 |
| `c4a-standard-64-lssd` `c4a-highmem-64-lssd` | 14 |
| `c4a-standard-72-lssd` `c4a-highmem-72-lssd` | 16 |
| *Only the [`-lssd` variants of the C4D machine types](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c4d_machine_types) support Local SSD.* ||
| `c4d-standard-8-lssd` `c4d-highmem-8-lssd` | 1 |
| `c4d-standard-16-lssd` `c4d-highmem-16-lssd` | 1 |
| `c4d-standard-32-lssd` `c4d-highmem-32-lssd` | 2 |
| `c4d-standard-48-lssd` `c4d-highmem-48-lssd` | 4 |
| `c4d-standard-64-lssd` `c4d-highmem-64-lssd` | 6 |
| `c4d-standard-96-lssd` `c4d-highmem-96-lssd` | 8 |
| `c4d-standard-192-lssd` `c4d-highmem-192-lssd` | 16 |
| `c4d-standard-384-lssd` `c4d-highmem-384-lssd` | 32 |
| C4N using Titanium SSD ||
| *Only the [`-lssd` variants of the C4N machine types](https://docs.cloud.google.com/compute/docs/network-optimized-machines.md.txt#c4n_series) support local Titanium SSD.* ||
| `c4n-highmem-4-lssd` `c4n-highmem-8-lssd` | 1 |
| `c4n-highmem-16-lssd` | 2 |
| `c4n-highmem-24-lssd` | 4 |
| `c4n-highmem-48-lssd` | 8 |
| `c4n-highmem-96-lssd` | 16 |
| `c4n-highmem-192-lssd` | 32 |
| C3 ||
| *Only the [`-lssd` variants of the C3 machine types](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c3_machine_types) support Local SSD.* ||
| `c3-standard-4-lssd` | 1 |
| `c3-standard-8-lssd` | 2 |
| `c3-standard-22-lssd` | 4 |
| `c3-standard-44-lssd` | 8 |
| `c3-standard-88-lssd` | 16 |
| `c3-standard-176-lssd` | 32 |
| C3D ||
| *Only the [`-lssd` variants of the C3D machine types](https://docs.cloud.google.com/compute/docs/general-purpose-machines.md.txt#c3d_machine_types) support Local SSD.* ||
| `c3d-standard-8-lssd` `c3d-highmem-8-lssd` | 1 |
| `c3d-standard-16-lssd` `c3d-highmem-16-lssd` | 1 |
| `c3d-standard-30-lssd` `c3d-highmem-30-lssd` | 2 |
| `c3d-standard-60-lssd` `c3d-highmem-60-lssd` | 4 |
| `c3d-standard-90-lssd` `c3d-highmem-90-lssd` | 8 |
| `c3d-standard-180-lssd` `c3d-highmem-180-lssd` | 16 |
| `c3d-standard-360-lssd` `c3d-highmem-360-lssd` | 32 |
| H4D using Titanium SSD ||
| `h4d-highmem-192-lssd` | 10 |
| `h4d-megamem-192-lssd` | 32 |
| Z4D using Titanium SSD |||
| *Each disk is 3,500 GiB in size:* ||
| `z4d-highmem-16-standardlssd` `z4d-highmem-8-highlssd` | 1 |
| `z4d-highmem-32-standardlssd` `z4d-highmem-16-highlssd` | 2 |
| `z4d-highmem-48-standardlssd` | 3 |
| `z4d-highmem-64-standardlssd` `z4d-highmem-32-highlssd` | 4 |
| `z4d-highmem-96-standardlssd` `z4d-highmem-48-highlssd` | 6 |
| `z4d-highmem-64-highlssd` | 8 |
| `z4d-highmem-192-standardlssd` `z4d-highmem-96-highlssd` | 12 |
| `z4d-highmem-384-standardlssd` `z4d-highmem-192-highlssd` | 24 |
| *Each disk is 7,000 GiB in size:* ||
| `z4d-highmem-384-standardlssd-metal` | 12 |
| Z4M using Titanium SSD |||
| *Each disk is 14,000 GiB in size:* ||
| `z4m-highmem-16` | 1 |
| `z4m-highmem-32` | 2 |
| `z4m-highmem-48` | 3 |
| `z4m-highmem-192` | 12 |
| Z3 using Titanium SSD |||
| *Each disk is 3,000 GiB in size:* ||
| `z3-highmem-8-highlssd` `z3-highmem-14-standardlssd` | 1 |
| `z3-highmem-16-highlssd` `z3-highmem-22-standardlssd` | 2 |
| `z3-highmem-22-highlssd` `z3-highmem-44-standardlssd` | 3 |
| `z3-highmem-32-highlssd` | 4 |
| `z3-highmem-44-highlssd` `z3-highmem-88-standardlssd` | 6 |
| `z3-highmem-88-highlssd` `z3-highmem-176-standardlssd` | 12 |
| *Each disk is 6,000 GiB in size:* ||
| `z3-highmem-192-highlssd-metal` | 12 |

### Machine types that require you to choose a number of Local SSD disks

The machine types listed in the following table don't attach Local SSD disks to
a newly created instance unless you specify how many disks to attach. Because
you can add Local SSD disks only during instance creation, use the information
in this section to determine how many Local SSD disks to attach when you create
an instance.

| Machine type | Number of Local SSD disks allowed per instance |
|---|---|
| N1 ||
| Machine types with T4 GPUs | 1 to 8, 16 |
| All other machine types | 1 to 8, 16, or 24 |
| N2 ||
| Machine types with 2 to 10 vCPUs, inclusive | 1, 2, 4, 8, 16, or 24 |
| Machine types with 12 to 20 vCPUs, inclusive | 2, 4, 8, 16, or 24 |
| Machine types with 22 to 40 vCPUs, inclusive | 4, 8, 16, or 24 |
| Machine types with 42 to 80 vCPUs, inclusive | 8, 16, or 24 |
| Machine types with 82 to 128 vCPUs, inclusive | 16 or 24 |
| N2D ||
| Machine types with 2 to 16 vCPUs, inclusive | 1, 2, 4, 8, 16, or 24 |
| Machine types with 32 or 48 vCPUs | 2, 4, 8, 16, or 24 |
| Machine types with 64 or 80 vCPUs | 4, 8, 16, or 24 |
| Machine types with 96 to 224 vCPUs, inclusive | 8, 16, or 24 |
| C2 ||
| Machine types with 4 or 8 vCPUs | 1, 2, 4, or 8 |
| Machine types with 16 vCPUs | 2, 4, or 8 |
| Machine types with 30 vCPUs | 4 or 8 |
| Machine types with 60 vCPUs | 8 |
| C2D ||
| Machine types with 2 to 16 vCPUs, inclusive | 1, 2, 4, 8 |
| Machine types with 32 vCPUs | 2, 4, 8 |
| Machine types with 56 vCPUs | 4, 8 |
| Machine types with 112 vCPUs | 8 |
| A2 Standard ||
| `a2-highgpu-1g` | 1, 2, 4, or 8 |
| `a2-highgpu-2g` | 2, 4, or 8 |
| `a2-highgpu-4g` | 4 or 8 |
| `a2-highgpu-8g` or `a2-megagpu-16g` | 8 |
| G2 ||
| `g2-standard-4` | 1 |
| `g2-standard-8` | 1 |
| `g2-standard-12` | 1 |
| `g2-standard-16` | 1 |
| `g2-standard-24` | 2 |
| `g2-standard-32` | 1 |
| `g2-standard-48` | 4 |
| `g2-standard-96` | 8 |
| G4 using Titanium SSD ||
| `g4-standard-6` | 0 |
| `g4-standard-12` | 1 |
| `g4-standard-24` | 2 |
| `g4-standard-48` | 4 |
| `g4-standard-96` | 8 |
| `g4-standard-192` | 16 |
| `g4-standard-384` | 32 |
| M1 ||
| `m1-ultramem-40` | 1 to 5 |
| `m1-ultramem-80` | 1 to 8 |
| `m1-megamem-96` | Not available |
| `m1-ultramem-160` | Not available |
| M3 ||
| `m3-ultramem-32` | 4, 8 |
| `m3-megamem-64` | 4, 8 |
| `m3-ultramem-64` | 4, 8 |
| `m3-megamem-128` | 8 |
| `m3-ultramem-128` | 8 |

## Reserve capacity for Local SSD disks

Reservations provide high assurance of capacity for zone-specific resources,
including Local SSD disks. You can use reservations to ensure that you have
Local SSD disks available when you need to use them for growth or disaster
recovery. For the different methods to reserve zone-specific resources in
Compute Engine, see
[Choose a reservation type](https://docs.cloud.google.com/compute/docs/instances/choose-reservation-type.md.txt).

For most machine series, reservations are also useful when you want to
receive [committed use discounts (CUDs) for your Local SSD disks](https://docs.cloud.google.com/compute/docs/disks/local-ssd.md.txt#lssd-cuds).

## Pricing

For each Local SSD disk you create, you are billed for the total capacity of
the disk for the lifetime of the instance that it is attached to.

For detailed information about Local SSD pricing and available discounts,
see [Local SSD pricing](https://docs.cloud.google.com/compute/disks-image-pricing#localssdpricing).

For Titanium SSD pricing, see the pricing for the machine series on the
[VM instance pricing](https://cloud.google.com/compute/vm-instance-pricing) page. For
example, for the pricing for Titanium SSD attached to a Z3 instance,
see the [Z3 instance pricing](https://docs.cloud.google.com/compute/vm-instance-pricing#section-4).

### Local SSD disks and Spot VM instances

If you start a [Spot VM](https://docs.cloud.google.com/compute/docs/instances/spot.md.txt) or preemptible VM
with a Local SSD disk, Compute Engine charges discounted
[spot prices](https://docs.cloud.google.com/compute/disks-image-pricing#localssdpricing)
for the Local SSD usage. Local SSD disks that are attached to Spot VMs or
preemptible VMs work like normal Local SSD disks, retain the same
[data persistence characteristics](https://docs.cloud.google.com/compute/docs/disks/local-ssd.md.txt#data_persistence),
and remain attached for the life of the VM.

Compute Engine doesn't charge you for Local SSD disk usage on a
Spot VM or preemptible VM if the VM is preempted within a minute after it
starts running.

### Committed use discounts for Local SSD disks

Resource-based commitments provide deep discounts for
Compute Engine resources in return for committing to using the resources
in a specific region for at least one year. You typically purchase commitments
for resources (vCPUs, memory, GPUs, and Local SSD disks) of a specific machine
series. When you use your resources, you receive qualifying resource usage at
discounted prices. To learn more about these discounts, see
[Resource-based committed use discounts](https://docs.cloud.google.com/compute/docs/instances/signing-up-committed-use-discounts.md.txt).

To purchase a commitment with Local SSD disks for most machine series, you
must also reserve the disks and attach the reservations to your commitment.
If you specify any local Titanium SSD disks in your commitment for use with
the following instances, then you don't need attached reservations for those
disks:

- C4
- C4A
- C4D
- C4N
- H4D
- Z3
- Z4D
- Z4M

For more information about attaching reservations to
commitments, see
[Attach reservations to resource-based commitments](https://docs.cloud.google.com/compute/docs/instances/reservations-with-commitments.md.txt#attach-reservations-to-resource-cuds).

## Use Local SSD disks with an instance

To use a Local SSD disk with a compute instance, you must complete the
following steps:

- [Add Local SSD disks when you create an instance](https://docs.cloud.google.com/compute/docs/disks/add-local-ssd.md.txt#create_local_ssd).
- [Format and mount Local SSD disks](https://docs.cloud.google.com/compute/docs/disks/add-local-ssd.md.txt#formatandmount) that you added to your instance.

### Device naming on Linux instances

The Linux device names for the disks attached to your instance depend on the
interface that you choose when creating the disks. When you use the `lsblk`
operating system command to view your disk devices, it displays the prefix
`nvme` for disks attached with the NVMe interface, and the prefix `sd` for
disks attached with the SCSI interface.

The ordering of the disk numbers or NVMe controllers is not predictable or
consistent across instance restarts. On the first boot, a disk might be
`nvme0n1` (or `sda` for SCSI). On the second boot, the device name for the same
disk might be `nvme2n1` or `nvme0n3` (or `sdc` for SCSI).

When accessing attached disks, you should use the symbolic links created in
`/dev/disk/by-id/` instead. These names persist across reboots.
For more information about symlinks, see
[Symbolic links for disks attached to an instance](https://docs.cloud.google.com/compute/docs/disks/disk-symlinks.md.txt).

For more information about device names, see
[Device naming on Linux instances](https://docs.cloud.google.com/compute/docs/disks/set-persistent-device-name-in-linux-vm.md.txt#device_names).

## Stop or suspend a Compute Engine instance with Local SSD disks

When you [stop](https://docs.cloud.google.com/compute/docs/instances/stop-start-instance.md.txt) or
[suspend an instance](https://docs.cloud.google.com/compute/docs/instances/suspend-resume-instance.md.txt),
Compute Engine discards the data of any
Local SSD disks attached to the instance by default. When you resume the
instance, all Local SSD disks attached to the instance are blank.

### Preserve Local SSD data when you stop or suspend an instance

> [!WARNING]
>
> **Preview
> --- Preserving Local SSD data**
>
>
> This feature is
>
> subject to the "Pre-GA Offerings Terms" in the General Service Terms section of the
> [Service Specific
> Terms](https://docs.cloud.google.com/terms/service-terms#1).
>
> Pre-GA features are available "as is" and might have limited support.
>
> For more information, see the
> [launch stage descriptions](https://cloud.google.com/products/#product-launch-stages).

When you stop or suspend a compute instance, you can optionally preserve the
data on the Local SSD disks attached to the instance.

When the stop or suspend operation starts, Compute Engine
performs a managed migration of the Local SSD disk data to durable storage.
When you resume or restart the instance, Compute Engine copies the
preserved data to Local SSD disks attached to the instance. After you resume or
restart the instance, you might have to
[remount the Local SSD disk into the file system](https://docs.cloud.google.com/compute/docs/disks/add-local-ssd.md.txt#formatandmount).

> [!WARNING]
> **Warning:** In the unlikely event that a hardware failure occurs while Compute Engine is copying the Local SSD data, Compute Engine attempts to preserve the Local SSD data. However, the data might be lost or corrupted when you restart or resume the instance if the attempt was unsuccessful.

You're billed for the storage space used to preserve the Local SSD data until
you restart or resume the instance. The used storage space consumes your
project's [`Persistent disk standard GB` quota](https://docs.cloud.google.com/compute/resource-usage.md.txt#disk_quota).

### Limitations

- Preserving Local SSD data is in [Preview](https://cloud.google.com/products/#product-launch-stages) only and is not covered under the GA terms for Compute Engine.
- Preserving Local SSD data isn't available for machine types that use Titanium SSD.
- You can't preserve the Local SSD data if you stop or suspend an instance that has more than 48 Local SSD disks attached.
- You can't preserve Local SSD data if you stop or suspend an instance from the Google Cloud console.
- Saving the Local SSD data begins only after the suspend or stop operation starts.
- Restoring the Local SSD data is a background process that begins after the instance starts. Reading data that isn't already restored triggers an immediate restoration of the requested data.
- If you're using Spot VMs or preemptible VMs and you opt to preserve Local SSD data during a suspend or stop operation, then the *Local SSD data is lost* if Compute Engine preempts the instance during the stop or suspend operation.

To learn how to preserve Local SSD data when you stop or suspend an instance,
see [Stop an instance with Local SSD disks](https://docs.cloud.google.com/compute/docs/instances/stop-start-instance.md.txt#stop-vm-local-ssd)
and
[Suspend an instance with Local SSD disks](https://docs.cloud.google.com/compute/docs/instances/suspend-resume-instance.md.txt#suspend-instance-local-ssd),
respectively.

## Delete Local SSD disks

To remove or delete Local SSD disks, you must delete the compute instance that
the disks are attached to. You can't delete Local SSD disks unless you delete
the instance.

Before you [delete a Compute Engine instance](https://docs.cloud.google.com/compute/docs/instances/deleting-instance.md.txt)
that has Local SSD disks attached, make sure that you migrate any critical data
on the Local SSD disks to a Persistent Disk, Hyperdisk, or to
another instance. Otherwise, the data on the Local SSD disks is
permanently lost.

## What's next

Learn how to [Create a compute instance with Local SSD disks](https://docs.cloud.google.com/compute/docs/disks/add-local-ssd.md.txt).