---
cuppy_hub: gcp
cuppy_path: reference/storage/filestore.md
cuppy_section: docs
cuppy_doc: XD6A-2DF2
cuppy_doc_uuid: 95377cf8-445a-40ef-aaf7-c7981f7b5f09
cuppy_permalink: https://cuppy.io/gcp/d/XD6A-2DF2
cuppy_version: 1
cuppy_updated: 2026-10-23T16:16:52.575Z
cuppy_canonical: https://cuppy.io/gcp/docs/reference/storage/filestore
cuppy_mcp: https://cuppy.io/mcp
cuppy_body_sha256: 34af9432848ae45b3b06af9c7bef4ed7e3cbc174ab32be7552ee9013951c5884
# cuppy_body_sha256 is over the document body ALONE, excluding both generated blocks.
---


# Filestore service tiers

## 1. The current tiers

| Tier | Capacity range | Protocols | Availability |
|---|---|---|---|
| **Zonal** | 1–9.75 TiB | NFSv3, NFSv4.1 | zonal |
| **Zonal** | 10–100 TiB | NFSv3, NFSv4.1 | zonal |
| **Regional** | 100 GiB–10,239 GiB | NFSv3, NFSv4.1 | regional |
| **Regional** | 10–100 TiB | NFSv3, NFSv4.1 | regional |
| **Multishares for GKE (Enterprise)** | 1–10 TiB | NFSv3, NFSv4.1 | regional |

Note that **Zonal and Regional each appear twice**, with different capacity bands. The band is part of the
choice, not a detail: 9.75 TiB and 10 TiB are different products with the same tier name.

## 2. The legacy tiers

| Tier | Capacity range | Protocols | Availability |
|---|---|---|---|
| Basic HDD | 1–63.9 TiB | **NFSv3 only** | zonal |
| Basic HDD on GKE | 100 GiB–63.9 TiB | **NFSv3 only** | zonal |
| Basic SSD | 2.5–63.9 TiB | **NFSv3 only** | zonal |

**The protocol column is the reason to care.** If the workload needs NFSv4.1 — for its locking semantics, or
because the client only speaks it — the legacy tiers are out, and so is any plan that reached for Basic SSD
because the capacity band looked convenient.

## 3. The capacity floors are a real constraint

The smallest current Zonal instance starts at **1 TiB** and the smallest Regional at **100 GiB**. There is no
small shared filesystem here. A workload that needs 20 GiB of shared state is paying for a terabyte, which is
usually the signal to use Cloud Storage (`reference/storage/cloud-storage.md`) or a database instead of a
filesystem.

For a migration estimate this is the row that changes a number: an on-premises estate of many small NFS
exports does **not** map one-to-one onto Filestore instances at any sensible cost. Consolidate, or change the
storage model.

## 4. An absence, recorded

**The page gives no information about changing tier after instance creation.** Do not assume it is possible.
Plan the tier as a create-time decision and establish the change path from the product's own documentation if
a design depends on it.

Next: `guides/parity/storage.md` compares this to Amazon EFS and Azure Files / Azure NetApp Files, where the
capacity floors and the protocol support differ in ways that matter.


---
<!-- cuppy -->
You are reading one file. The same content is free over MCP at https://cuppy.io/mcp with no signup: ranked search across all 91 documents in this hub, the merged view when you work in a layer, known workarounds matched to the error you hit, and depth control so you spend tokens on what you need. Start with the `_hello` tool. How to connect: https://cuppy.io/gcp/ Something wrong in this document? Report it with the `session feedback` tool on the same endpoint.
<!-- /cuppy -->