Cloud Storage FUSE와 같은 객체 스토리지와 비교할 때, 이 제품은 일부 파일 시스템 시맨틱스를 제공하지만 Filestore에서 제공하는 파일 스토리지의 일부 더 강력한 특성이 없습니다. Cloud Storage FUSE가 아닌 Filestore에서 지원하는 기능의 예는 다음과 같습니다.
Filestore 인스턴스는 공유 VPC 네트워크를 포함하여 동일한 VPC 네트워크에 있는 모든 클라이언트에 연결할 수 있습니다. Cloud VPN 또는 Cloud Interconnect를 사용하여 온프레미스 머신과 같은 원격 네트워크의 클라이언트에 연결할 수도 있습니다.
클라이언트의 IP 주소를 기반으로 Filestore 인스턴스 데이터에 대한 클라이언트의 액세스 수준을 제어할 수 있습니다.
인스턴스에 대한 IP 기반 액세스 제어 규칙은 인스턴스 생성 중 및 생성 후에 만들거나 수정할 수 있습니다.
IAM 권한 및 역할을 사용하여 Filestore 리소스를 생성, 수정, 확인할 수 있는 Google Cloud 사용자를 제어할 수도 있습니다.
데이터 복구 옵션
다음 섹션에서는 Filestore 인스턴스 데이터 복구 옵션에 대해 설명합니다.
백업
Filestore 백업은 모든 사용자 데이터 및 일부 인스턴스 메타데이터를 포함하는 Filestore 인스턴스의 특정 시점 사본입니다. 어느 리전에서든 인스턴스의 백업을 만든 다음 이를 사용하여 어느 리전에서든 인스턴스를 기존 Filestore 인스턴스 또는 새 인스턴스로 복원할 수 있습니다.
스냅샷
Filestore 스냅샷은 스냅샷이 만들어진 시점의 Filestore 인스턴스 데이터 상태를 유지합니다.
스냅샷을 사용하여 개별 파일이나 디렉터리를 복원하거나 인스턴스를 스냅샷 상태로 완전히 되돌릴 수 있습니다.
신뢰성
다음 섹션에서는 Filestore 인스턴스 신뢰성에 대해 설명합니다.
영역 인스턴스
기본 및 영역 등급 Filestore 인스턴스는 기기 오류로부터 데이터를 보호하기 위해 영역 내 스토리지 중복성을 제공하는 영역별 리소스입니다. 하지만 전력 중단 또는 데이터 센터 유지보수 등을 이유로 영역 작동이 중지되면 영역 작동 중지 기간 동안 해당 영역에 있는 인스턴스를 사용할 수 없게 됩니다.
리전에 하나 이상의 영역 오류가 있는 경우에도 준비되고 실행되는 모든 영역에 기본 및 영역 등급 인스턴스를 만들 수 있습니다.
리전 인스턴스
지역 및 엔터프라이즈 등급 Filestore 인스턴스는 리전별 리소스입니다.
영역 오류가 발생할 경우 이러한 인스턴스는 계속해서 데이터를 제공하고 새 쓰기를 허용하므로 영역 오류가 클라이언트에 투명하게 처리됩니다. 또한 Filestore는 NFS에 필요한 엄격한 일관성 정책을 적용합니다. 클라이언트가 데이터를 쓸 경우, Filestore는 영역 오류가 발생해도 후속 읽기에서 올바른 데이터를 반환하도록 변경이 유지될 때까지 확인을 반환하지 않습니다.
영역 오류가 발생하면 Google Cloudconsole 또는 Filestore API를 통한 작업을 몇 시간 동안 사용하지 못할 수 있습니다.
이러한 인스턴스에는 NFS 데이터 액세스 중단이 발생하지 않지만 영역이 복구될 때까지 성능이 저하될 수 있습니다. 또한 영역 장애가 발생한 리전에서는 지역 또는 엔터프라이즈 인스턴스를 만들 수 없습니다.
[[["이해하기 쉬움","easyToUnderstand","thumb-up"],["문제가 해결됨","solvedMyProblem","thumb-up"],["기타","otherUp","thumb-up"]],[["이해하기 어려움","hardToUnderstand","thumb-down"],["잘못된 정보 또는 샘플 코드","incorrectInformationOrSampleCode","thumb-down"],["필요한 정보/샘플이 없음","missingTheInformationSamplesINeed","thumb-down"],["번역 문제","translationIssue","thumb-down"],["기타","otherDown","thumb-down"]],["최종 업데이트: 2025-09-04(UTC)"],[[["\u003cp\u003eFilestore is a fully managed file server on Google Cloud that supports connections from Compute Engine VMs, Google Kubernetes Engine (GKE) clusters, external datastores, on-premises machines, and Cloud Run services, providing a versatile solution for various client types.\u003c/p\u003e\n"],["\u003cp\u003eFilestore allows multiple concurrent application instances to access the same file system simultaneously, offering multiple reader, multiple writer access, which is particularly beneficial for GKE users.\u003c/p\u003e\n"],["\u003cp\u003eFilestore provides several service tiers, including Basic, Regional, Enterprise, and Zonal, each designed for specific use cases such as file sharing, mission-critical workloads, and high-performance computing.\u003c/p\u003e\n"],["\u003cp\u003eFilestore supports NFSv3 and NFSv4.1 protocols, each best suited for different use cases, and offers data protection through encryption at rest and in transit, along with IP-based access control and IAM permissions.\u003c/p\u003e\n"],["\u003cp\u003eFilestore provides data recovery options like backups and snapshots, enabling users to restore data to any region or revert instances to a previous state, and regional or enterprise-tier instances offer high availability in the event of zone failures.\u003c/p\u003e\n"]]],[],null,["# Filestore overview\n\nFilestore instances are fully managed file servers on Google Cloud that\ncan be connected to a number of client types:\n\n- [Compute Engine VMs](/filestore/docs/mounting-fileshares)\n- [Google Kubernetes Engine (GKE) clusters](/filestore/docs/filestore-for-gke)\n- [External datastores such as Google Cloud VMware Engine](/filestore/docs/gcve-datastores)\n- [On-premises machines](/filestore/docs/remote-mounting)\n- [Cloud Run services](/filestore/docs/mount-filestore-cloud-run)\n\nOnce provisioned, you can scale the capacity of your instances according to need\nwithout any downtime.\n\nWhy Filestore?\n--------------\n\nGoogle Cloud offers three main types of data storage: block, file, and object\nstorage.\n\nAs a type of persistent file storage, Filestore supports multiple\nconcurrent application instances accessing the same file system simultaneously.\n\nFor Google Kubernetes Engine users, for example, Filestore provides multiple reader,\nmultiple writer access, letting you mount your GKE\nPersistentVolumes as read-write by many nodes.\n\nFilestore offers a versatile alternative to block storage products,\nsuch as Persistent Disk, which support only\n[limited options for multiple writer access to a disk](/compute/docs/disks/sharing-disks-between-vms).\n\nWhen compared to object storage, such as Cloud Storage FUSE, while that product does\noffer some file system [semantics](https://github.com/GoogleCloudPlatform/gcsfuse/blob/master/docs/semantics.md),\nit [lacks some of the more robust characteristics of file storage](/storage/docs/gcs-fuse#expandable-1)\nprovided by Filestore. Some examples of capabilities supported by\nFilestore and not Cloud Storage FUSE include the following:\n\n- POSIX compliance\n- Hard links and file locking\n- Concurrency control for multiple writes on the same object\n\nFor more information, see the following resources:\n\n- [Compare the relative advantages of block, file, and object storage](/architecture/storage-advisor#review_the_storage_options).\n- [Review the storage options for HPC workloads in Google Cloud](/architecture/parallel-file-systems-for-hpc#storage-options-for-hpc-workloads-in-google-cloud).\n- [Learn more about Filestore support for GKE](/filestore/docs/filestore-for-gke).\n- [Learn more about the limitations of Cloud Storage FUSE](/storage/docs/gcs-fuse#expandable-1).\n- [Learn about storage options and access modes for GKE clusters](/kubernetes-engine/docs/concepts/storage-overview).\n\nService tiers\n-------------\n\nFilestore offers multiple service tiers that vary in capacity,\nperformance, and features. Each service tier is tailored for specific use cases:\n\n- **Basic tier:** File sharing, software development, web hosting, basic AI.\n- **Zonal tier:** HPC, batch compute, EDA, media rendering and transcoding, advanced AI, large data sets.\n- **Regional tier:** Mission-critical workloads requiring high availability.\n- **Enterprise tier:** Mission-critical workloads requiring high availability.\n\nFor details, see [Service tiers](/filestore/docs/service-tiers).\n\nProtocol support\n----------------\n\nFilestore supports the following file system protocols:\n\n- NFSv3\n\n - Available in all service tiers\n - Supports bidirectional communication between the client and server\n- NFSv4.1\n\n - Available in [zonal, regional, and enterprise service tiers](/filestore/docs/service-tiers)\n - Supports client and server authentication, message integrity checks, and in-transit data encryption\n\nEach protocol is best suited to specific use cases.\n\nFor help understanding which protocol may be right for you, see\n[About supported protocols](/filestore/docs/about-supported-protocols).\n\nConnectivity\n------------\n\nFilestore instances can connect to any clients that are on the same VPC\nnetwork, including [Shared VPC networks](/vpc/docs/shared-vpc). You can\nalso connect to clients on remote networks, such as an on-premises machine,\nusing\n[Cloud VPN](/network-connectivity/docs/vpn/concepts/overview) or\n[Cloud Interconnect](/network-connectivity/docs/interconnect/concepts/overview).\n\nNetworking\n----------\n\nFor information related to Filestore networking requirements, see the\nfollowing resources:\n\n- [Filestore networking and IP resource requirements](/filestore/docs/network-ip-requirements)\n- [Configure NFS ports on client VMs](/filestore/docs/setting-nfs-ports)\n- [Configure firewall rules](/filestore/docs/configuring-firewall)\n- [Create an instance on a Shared VPC network in service projects](/filestore/docs/shared-vpc)\n- [Supported file system protocols](/filestore/docs/about-supported-protocols)\n\nPerformance\n-----------\n\n[Custom performance](/filestore/docs/custom-performance) lets you define the performance settings for\nyour Filestore instances independently of the specified capacity.\nThis means you can optimize your Filestore instances for your workload\nrequirements and scale your file shares to meet the demands of your applications\nwithout resizing the capacity.\n| **Note:** Custom performance is available for regional and zonal [service tiers](/filestore/docs/service-tiers).\n\nFor details on performance settings, limits, and testing see [performance](/filestore/docs/performance).\n\nData protection\n---------------\n\nThe following sections discuss Filestore instance data protection.\n\n### Encryption at rest\n\nBy default, Filestore automatically encrypts your data at rest. The\ndurable storage behind each Filestore instance is encrypted with\nsystem-defined keys that are managed by Google.\n\nWhen you delete a Filestore instance, Google discards the encryption\ninformation used by the instance, rendering the data irretrievable as per the\ndescription in [Data deletion on Google Cloud](/security/deletion).\n\nIf you need more control over the keys that protect your data, you can also use\n[customer-managed encryption keys (CMEK)](/filestore/docs/cmek) with Filestore.\n\nFor details, see\n[Encryption at rest in Google Cloud](/security/encryption/default-encryption).\n\n### Encryption in transit\n\nAlthough NFSv3 does not encrypt data in transit, all in-transit data to and\nwithin Google Cloud is encrypted.\n\nFor details, see\n[Encryption in Transit in Google Cloud](/security/encryption-in-transit).\n\n### Access control\n\nYou can control the level of access that a client has on Filestore\ninstance data based on the client's IP address.\n[IP-based access control](/filestore/docs/access-control#ip-based_access_control)\nrules for an instance can be created or modified during and after instance\ncreation.\n\nYou can also control which Google Cloud users can create, edit, and view\nFilestore resources by using [IAM permissions and roles](/filestore/docs/iam).\n\nData recovery options\n---------------------\n\nThe following sections discuss Filestore instance data recovery options.\n\n### Backups\n\n[Filestore backups](/filestore/docs/backups) are point-in-time copies of\na Filestore instance that includes all user data and some instance\nmetadata. You can create a backup of an instance in any region and then use it\nto restore the instance in any region to an existing Filestore instance\nor a new instance.\n\n### Snapshots\n\nA [Filestore snapshot](/filestore/docs/snapshots) preserves the state of\nyour Filestore instance data at the time that the snapshot is created.\nYou can use snapshots to restore individual files or directories or completely\nrevert your instance to the state of a snapshot.\n\nReliability\n-----------\n\nThe following sections discuss Filestore instance reliability.\n\n### Zonal instances\n\nBasic- and zonal-tier Filestore instances are zonal resources that\nfeature in-zone storage redundancy to protect your data against equipment\nfailure. However, if a zone goes down due to an outage or data center\nmaintenance, the instances that reside in that zone become unavailable for the\nduration that the zone is down.\n\nYou can create basic- and zonal-tier instances to any zone that's up and\nrunning even if there's one or more zone failures in the region.\n\n### Regional instances\n\nRegional and enterprise-tier Filestore instances are regional resources.\nIn the event of a zone failure, these instances continue to serve data and accept\nnew writes, making the zone failure transparent to clients. Also,\nFilestore adopts the strict consistency policy required by NFS. When a\nclient writes data, Filestore doesn't return an acknowledgment until the\nchange is persisted so that subsequent reads return the correct data, even\nduring a zone failure.\n\nDuring a zone failure, operations through the Google Cloud\nconsole or to the Filestore API may be unavailable for a few hours.\nThese instances don't experience NFS data access interruptions, but you may\nexperience some performance degradation until the zone recovers. Also, you can't\ncreate a regional or enterprise instance in a region experiencing zone failures.\n\n### Zone failure identification\n\nYou can check for zone failures on the\n[Google Cloud Status Dashboard](https://status.cloud.google.com).\n\nWhat's next\n-----------\n\n- Learn more about Filestore's [service tiers](/filestore/docs/service-tiers) to decide which service tier is right for you.\n- See the [regions that support Filestore](/filestore/docs/regions).\n- [Compare the relative advantages of block, file, and object\n storage](/architecture/storage-advisor#review_the_storage_options).\n- [Review the storage options for HPC workloads in\n Google Cloud](/architecture/parallel-file-systems-for-hpc#storage-options-for-hpc-workloads-in-google-cloud)."]]