Configuring replication for use with Multi-Region Access Points
When you make a request to a Multi-Region Access Point endpoint, Amazon S3 automatically routes the request to the
bucket that is closest to you. Amazon S3 doesn't consider the contents of the request when making
this decision. If you make a request to GET
an object, your request might be
routed to a bucket that doesn't have a copy of this object. If that happens, you receive an
HTTP status code 404 (Not Found) error. For more information about Multi-Region Access Point request routing,
see Multi-Region Access Point request routing.
If you want the Multi-Region Access Point to be able to retrieve the object regardless of which bucket receives the request, you must configure Amazon S3 Cross-Region Replication (CRR).
For example, consider a Multi-Region Access Point with three buckets:
-
A bucket named
my-bucket-usw2
in the Regionus-west-2
that contains the objectmy-image.jpg
-
A bucket named
my-bucket-aps1
in the Regionap-south-1
that contains the objectmy-image.jpg
-
A bucket named
my-bucket-euc1
in the Regioneu-central-1
that doesn't contain the objectmy-image.jpg
In this situation, if you make a GetObject
request for the object
my-image.jpg
, the success of that request depends upon which bucket
receives your request. Because Amazon S3 doesn't consider the contents of the request, it might
route your GetObject
request to the my-bucket-euc1
bucket if that
bucket responds of closest proximity. Even though your object is in a bucket in the Multi-Region Access Point,
you will get an HTTP 404 Not Found error because the individual bucket that received your
request didn't have the object.
Enabling Cross-Region Replication (CRR) helps avoid this result. With appropriate
replication rules, the my-image.jpg
object is copied over to the
my-bucket-euc1
bucket. Therefore, if Amazon S3 routes your request to that
bucket, you can now retrieve the object.
Replication works as normal with buckets that are assigned to a Multi-Region Access Point. Amazon S3 doesn't perform any special replication handling with buckets that are in Multi-Region Access Points. For more information about configuring replication in your buckets, see Setting up live replication overview.
Recommendations for using replication with Multi-Region Access Points
For the best replication performance when working with Multi-Region Access Points, we recommend the following:
-
Configure S3 Replication Time Control (S3 RTC). To replicate your data across different Regions within a predictable time frame, you can use S3 RTC. S3 RTC replicates 99.99 percent of new objects stored in Amazon S3 within 15 minutes (backed by a service-level agreement). For more information, see Meeting compliance requirements with S3 Replication Time Control. There are additional charges for S3 RTC. For information, see Amazon S3 pricing
. -
Use two-way (bidirectional) replication to support keeping buckets synchronized when buckets are updated through the Multi-Region Access Point. For more information, see Create two-way replication rules for your Multi-Region Access Point.
-
Create cross-account Multi-Region Access Points to replicate data to buckets in separate Amazon Web Services accounts. This approach provides account-level separation, so that data can be accessed from and replicated across different accounts in different Regions other than the source bucket. Setting up cross-account Multi-Region Access Points comes at no additional cost. If you're a bucket owner but don't own the Multi-Region Access Point, you pay only for data transfer and request costs. Multi-Region Access Point owners pay for data routing and internet-acceleration costs. For more information, see Amazon S3 pricing
. -
Enable replica modification sync for each replication rule to also keep metadata changes to your objects in sync. For more information, see Enabling replica modification sync.
-
Enable Amazon CloudWatch metrics to monitor replication events. CloudWatch metrics fees apply. For more information, see Amazon CloudWatch pricing
.