[[["易于理解","easyToUnderstand","thumb-up"],["解决了我的问题","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["很难理解","hardToUnderstand","thumb-down"],["信息或示例代码不正确","incorrectInformationOrSampleCode","thumb-down"],["没有我需要的信息/示例","missingTheInformationSamplesINeed","thumb-down"],["翻译问题","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["最后更新时间 (UTC):2025-09-08。"],[],[],null,["# About instance replication\n\nFilestore now offers asynchronous replication for your instances.\n\nA source instance can be continuously and asynchronously replicated to a\nstandby instance in a location of your choice.\n\nTo select customers, Filestore offers instance replication support for\ninstances created in the following\n[service tiers](/filestore/docs/service-tiers):\n\n- Zonal\n- Regional\n- Enterprise\n\nInstance replication compared to other data recovery options\n------------------------------------------------------------\n\nThe following sections cover the advantages of instance replication compared to\nsnapshots and backups.\n\n### Snapshots\n\n*[Snapshots](/filestore/docs/snapshots)* are resources that consume capacity\nwithin the instance and allow you to revert the current state of your instance\ndata to a previous, specific point in time. Users can also elect to revert to a\nprevious version of a specific file.\n\nSnapshots don't replicate any data or consume capacity until the data on the\ninstance is modified. All snapshots of an instance share in-common data, meaning\nthat the instance preserves only the differences between the snapshots.\n\nWhile snapshots do offer cost efficiency when compared to other Filestore\ndata recovery operations, available instance capacity continuously decreases as\nfile modifications are made.\n\nReverting an instance to a previous state is destructive, as it deletes the\nlatest version of instance data, and should be used carefully.\n\n### Backups\n\n*[Backups](/filestore/docs/backups)* are external resources that reside outside\nof the instance, consuming their own separate capacity. The first backup is a\ncomplete copy of instance data, with each subsequent backup only consuming the\ndata necessary to track [incremental and differential](/filestore/docs/backups#backup-creation)\nchanges since the previous backup. Internally, a [backup chain's](/filestore/docs/backups#backup-creation)\nhistory is tracked using snapshots, which consume capacity on the source\ninstance.\n\n### Instance replication\n\n*Instance replication* pairs a source instance with a replica instance,\na separate resource in a secondary location that continuously monitors any\nchanges made to the source and asynchronously replicates those changes to the\nreplica instance within a target [recovery point objective (RPO)](/architecture/dr-scenarios-planning-guide#basics_of_dr_planning)\nof approximately 30 minutes.\n\nThis process relies on snapshots and similarly consumes capacity. As snapshots\nbecome obsolete, they are deleted to free up instance capacity in both instances.\n\nThe replica instance is a continuously updated, complete copy of the source\ninstance. Most copies are scheduled every five to ten minutes. Metrics are\navailable indicating the timestamp of the last full instance copy. For more\ninformation, see [Monitoring](#monitoring).\n\nInstance pair roles\n-------------------\n\nWhen instance replication is enabled, each instance in the pair is assigned a\nrole:\n\n- **`ACTIVE`**\n\n The source instance.\n- **`STANDBY`**\n\n The replica instance.\n\nRole changes are not automatic and can only be initiated by the user.\n\nPromote replica\n---------------\n\nWhile the replica instance can't be mounted or written to directly, in the event\nof a service outage, the replica can perform a\n[`promote-replica`](/sdk/gcloud/reference/filestore/instances/promote-replica)\noperation. This same operation is accessible from the Google Cloud console.\n\nThis operation initiates the following actions:\n\n- Stops replication between the `ACTIVE` and `STANDBY` instances.\n- Unpairs the two resources.\n- Promotes the replica to a regular instance that is capable of client connections and writes just as any other instance.\n- The replication role, whether `ACTIVE` or `STANDBY`, is removed from both instances.\n- When the operation successfully completes, the replica instance's state changes from `PROMOTING` to `READY`.\n\nOnce your application is back online, this new source instance can then be\npaired with a new replica and you can initiate instance replication once again.\n\nFor more information, see [Break replication and promote replica](/filestore/docs/configure-instance-replication#break-replication).\n\n### Failback\n\nThe `promote-replica` operation initiates a break in replication. It is a\nswitchover not a failover event, meaning that [failback](https://en.wikipedia.org/wiki/Failover)\nisn't automatic. Once the operation is complete, administrators must\nreconnect their applications to the new source instance.\n\nLimitations\n-----------\n\nThe following limitations apply:\n\n- Not available in the following service tiers:\n\n - Basic HDD\n - Basic SSD\n- The following specifications must be the same for each instance in the pair:\n\n - Project\n - [Service tier](/filestore/docs/service-tiers), including capacity range\n\nThe following sections cover other feature limitations related to RPO, operations,\nnetwork traffic, performance, and availability in detail:\n\n### RPO\n\n- Quoted RPO times are approximate and not supported under any\n [Filestore service level agreement (SLA)](/filestore/sla).\n\n- The target 30-minute RPO applies to instances with a change rate of\n 100 MB per second and an IOPS rate of 300 per second, where IOPS are\n defined as any `create`, `edit`, or `delete` operation applied to any file or\n directory.\n\n Instances with a higher change rate may experience longer replication periods.\n RPO times vary and don't strictly map linearly with higher change rates. For\n example, if the change rate doubles, the RPO window won't necessarily double.\n- Instance replication is specified at instance creation. If you want to pair a\n replica with a source instance, you must enable instance replication on the\n instance when creating it, and then pair it with the replica. An existing\n instance can't be used as a replica.\n\n### Operations\n\n- The following operations are disabled on the `ACTIVE` instance:\n\n - Instance deletion\n - Revert to snapshot\n- The following limitations apply to the `STANDBY` instance:\n\n - Doesn't support NFS access\n - Backups are disabled\n - Snapshots are disabled\n\n### Network traffic\n\n- Continuous data replication between the source and replica instances accrues network traffic fees. For more information, see [Filestore pricing](/filestore/pricing) or your Filestore representative.\n\n### Performance\n\n- Continuous data replication between the pair impacts IOPS performance on the source instance.\n\n### Availability\n\n- Some restrictions apply when choosing a region for your replica instance. For\n more information, see your Filestore representative.\n\n- Deleting the source instance won't delete its associated replica. If the\n source instance no longer exists, and you want to delete the instance\n previously used as its replica, use [`gcloud` to locate and delete the instance](/filestore/docs/configure-instance-replication#gcloud-delete).\n\nRecommendations\n---------------\n\nFor a robust disaster recovery solution, users are strongly encouraged to choose\na separate region rather than a separate zone for the replica instance.\n\nMonitoring\n----------\n\nThe `time_since_last_replication` metric represents the timestamp of the last\nfull copy of the active instance.\n\nFor more information, see the following resources:\n\n- [Monitor instance replication](/filestore/docs/configure-instance-replication#monitor-rpo)\n- [Monitoring instances](/filestore/docs/monitoring-instances)\n- [Filestore metrics](/monitoring/api/metrics_gcp?_ga=2.63538607.-1445436237.1687238861#gcp-file)\n\nPricing\n-------\n\nWhen you use instance replication, charges for instances and cross-region\nnetworking apply. Replicating instances is offered at no additional charge. The\nreplica instance is a continuously updated, complete copy of the source\ninstance. Pricing is commensurate with the data recovery support provided in a\ncontinuous data recovery option suitable for enterprise workloads.\n\nFor more information, see your Filestore representative.\n\nWhat's next\n-----------\n\n- [Configure instance replication](/filestore/docs/configure-instance-replication)\n- [Instance replication reference documentation](/filestore/docs/reference/rest/v1/projects.locations.instances/promoteReplica)"]]