Undocumented Amazon S3 APIs and third-party extensions. This documents features outside the official Amazon S3 REST APIs. Note that these are not consistently supported across implementations.
Object metadata exposes the number of multipart upload parts via the
x-amz-mp-parts-count
header. Subsequent GET object requests can retrieve
individual parts via the partNumber
parameter. This allows retrieving
individual multipart upload parts, including the original ETag.
HEAD bucket exposes the region via x-amz-bucket-region
even if the request
lacks authentication. This allows clients to interrogate a bucket so that they
can route subsequent requests to the correct endpoint.
Single part upload ETag are the MD5 hash of the object content. Multipart upload ETag concatenate the MD5 of each part, MD5 this data, and append a hyphen and the number of parts.
allow-unordered=true
parameter overrides key sorting which allows higher
performance with large buckets. Cannot be used with delimiter.
Returns bucket statistics via x-rgw-bytes-used
and x-rgw-object-count
headers.
Append to an existing object at a given offset via the append
and position
parameters. Responses include x-oss-next-append-position
and
x-rgw-next-append-position
which indicate the next valid append offset.
If-Match
and If-None-Match
(Ceph) and x-goog-if-generation-match
and
x-goog-if-metageneration-match
(Google) headers allow atomic replacement of
objects.
Initiate a resumable upload with POST object with the x-goog-resumable
header. Use the response Location
to issue a PUT object with given
Content-Length and Content-Range headers. Continuing issuing these requests to
build larger objects.
Allows interaction with ZIP file contents as if they were objects. HeadObject,
GetObject, and ListObjectsV2 support these operations when the header
x-minio-extract
has a true
value.