# \[doc\] \[S3\] Pydio creates a \*new\* container suffixed by \`+segments\`
**URL:** https://forum.pydio.com/t/doc-s3-pydio-creates-a-new-container-suffixed-by-segments/3121
**Category:** Pydio Cells
**Tags:** install, s3
**Created:** [March 3, 2020, 3:34pm UTC](https://forum.pydio.com/t/doc-s3-pydio-creates-a-new-container-suffixed-by-segments/3121 "2020-03-03T15:34:25Z")
**Posts on this page:** 4
**Page:** 1
### Author:  [@drzraf](https://forum.pydio.com/u/drzraf)
#### Post date: [March 3, 2020, 3:34pm UTC](https://forum.pydio.com/t/doc-s3-pydio-creates-a-new-container-suffixed-by-segments/3121/1 "2020-03-03T15:34:25Z")
While I provided an existing S3-compatible container to my instance during the initial configuration, I got surprised to see that a brand new container got seemingly automatically created by Pydio.
Its name is derived from the original one by appending `+segments`.
There is no information about this at [https://pydio.com/en/docs/cells/v2/s3-compatible-storage](https://pydio.com/en/docs/cells/v2/s3-compatible-storage) and I find this quite surprising.
What would happen, for example, if the access token provided did not allow for new container creation (which would be safer configuration)?
---
### Author:  [@drzraf](https://forum.pydio.com/u/drzraf)
#### Post date: [April 28, 2020, 2:40pm UTC](https://forum.pydio.com/t/doc-s3-pydio-creates-a-new-container-suffixed-by-segments/3121/2 "2020-04-28T14:40:57Z")
Reason is the handling of multi-part uploads by Swift S3 middleware: [https://docs.openstack.org/swift/latest/middleware.html#bucket-segments](https://docs.openstack.org/swift/latest/middleware.html#bucket-segments)
---
### Author:  [@charles](https://forum.pydio.com/u/charles)
#### Post date: [June 24, 2020, 12:11pm UTC](https://forum.pydio.com/t/doc-s3-pydio-creates-a-new-container-suffixed-by-segments/3121/3 "2020-06-24T12:11:15Z")
So nothing to do with Pydio, right?
That said, how is the SwiftS3 gateway behaving? Good ?
---
### Author:  [@drzraf](https://forum.pydio.com/u/drzraf)
#### Post date: [June 24, 2020, 1:36pm UTC](https://forum.pydio.com/t/doc-s3-pydio-creates-a-new-container-suffixed-by-segments/3121/4 "2020-06-24T13:36:20Z")
- About `+segments`: Nothing to do with Pydio indeed
- About Swift-backed storage, I wasn’t satisfied (at least in 2.0.5) but I can’t neither blame Pydio nor the middleware for that. Basically my guess is that most file operations (upoad/copy/move/…) are implemented using file-system primitives instead of relying on built-in Swift API calls. This causes many negative effects.
But now I want to see if Pydio could be usable by users to browse/download/tag files uploaded (unencrypted) straight into the object-storage using Cyberduck as a Swift-enabled client to bypass performances/stability bottlenecks). This relies on the assumption that the `resync` operation be reliable and efficient.
About encryption: Swift provides encrypted-at-rest, but it’s not implemented by my provider.