S3 API Usage Guidelines
The following table describes usage guidelines for naming or mapping S3 API bucket
directories and objects.
Table 2 Usage Guidelines: S3 API
|
|
|
1
|
The S3 API creates sub-directories in bucket for each
"/" it encounters in the object's
name.
For example, when the user uploads an object called
"accounting/billing.pdf", a directory called
"accounting" is created in the bucket.
|
|
2
|
Objects that are mapped to the Oracle ZFS Storage Appliance
directories, such as an object with names that end with
"/", cannot store content for themselves.
However, content can be created in them.
For example: object "foo/" cannot have any data
associated with it but object "foo/bar"
can.
|
|
3
|
Any file name not permitted by the Oracle ZFS Storage Appliance
filesystem is also not permitted while creating objects in the S3
API.
For example, objects names with any of the following characters
are not permitted: "//", "/",
".". Object names containing double
"//" are also not permitted
(foo//bar).
|
|
4
|
Object names that are relative paths are not permitted.
For example, "../../../foo"
|
|
5
|
Oracle ZFS Storage Appliance limits file and directory names to
255 characters. The same character limit (255) applies to the
virtual directory names and object names for the Oracle ZFS Storage
Appliance S3 API.
|
|
6
|
When bucket versioning is either enabled or suspended, take the
following naming conventions into consideration:
-
When versioning is enabled for a bucket, previous versions
are saved in the same directory as the current version of
the object.
-
The current version of an object is renamed when a new
object or delete marker takes its place. The new name of the
current version follows the pattern: <object
name>-<versionId>. For
example, billing.pdf-0001.
|
|