This commit fixes the issue #47856. When you upload a file into a group folder and when you use a single S3 bucket as primary storage, the final move operation hangs for a long time. In the background, Nextcloud initiates a copy-delete sequence from the bucket into the bucket, with causes a lot unnecessary overhead. Nextcloud thinks that the file must be imported to another storage and does not recognize that everything is done on the same object bucket. In that case, the import step can be completely skipped, which saves time, network bandwidth and reduces the load on the object storage.
The behavior improves a lot with https://github.com/nextcloud/server/pull/46013. However, there are still some put messages that are being sent to the object storage when you use an object storage as primary storage and upload files into a group folder.
Co-authored-by: Kate <26026535+provokateurin@users.noreply.github.com>
Signed-off-by: Christoph Fiehe <c.fiehe@eurodata.de>
When copying or moving between two local storages the source path (on
disk) to copy or move from is got from the unjailed path of the source
storage. However, if the source storage has more than one jail getting
the unjailed path resolves the most external jail, but the source path
needs to be got from the most internal jail instead (the one closer to
the local storage).
This can happen, for example, with a shared groupfolder: in that case
there is an external jail for the shared storage, and one internal jail
for the groupfolder storage wrapped by the shared storage.
Signed-off-by: Daniel Calviño Sánchez <danxuliu@gmail.com>
Storages that do not have a dedicated owner (e.g. groupfolders, external
storages) currently always assume the current session user as the owner.
This leads to several issues when there is no user session but a node is
obtained through a user folder.
In order to have the correct user available we need to pass the user
that is used to setup a mountpoint along to the storage layer as we
generally assume that an owner is available for those.
Signed-off-by: Julius Härtl <jus@bitgrid.net>
There is no strong requirement to have only one mount for a given
storage id. Also the error in this case would be misleading.
Signed-off-by: Côme Chilliet <come.chilliet@nextcloud.com>
The owner was not set on the file info causing e.g. webdav searches to never return the known owner.
Signed-off-by: Ferdinand Thiessen <opensource@fthiessen.de>
If a mountpoint is returned from a `View` the name must match the view scoped name
and not the global name.
For example group folders have a name like `1` or `2` (the ID),
but the name of the FileInfo returned fro mthe View should be the mount point name,
like `First groupfolder` or similar.
Signed-off-by: Ferdinand Thiessen <opensource@fthiessen.de>
Fixes#45637
The support for s3-accelerate added in #44496 introduced a regression in AWS S3 environments when `hostname` is blank (which is a valid configuration w/ AWS since the hostname gets auto-generated).
Signed-off-by: Josh <josh.t.richards@gmail.com>
