Skip to content

Commit

Permalink
Merge pull request #10740 from nextcloud/dev/chunking-v2
Browse files Browse the repository at this point in the history
  • Loading branch information
juliusknorr authored Aug 10, 2023
2 parents 9cdf815 + d2b012c commit 6d3a740
Showing 1 changed file with 103 additions and 3 deletions.
106 changes: 103 additions & 3 deletions developer_manual/client_apis/WebDAV/chunking.rst
Original file line number Diff line number Diff line change
Expand Up @@ -10,13 +10,113 @@ Uploading large files is always a bit problematic as your connection can be inte
which will fail your entire upload. Nextcloud has a chunking API where you can
upload smaller chunks which will be assembled on the server once they are all uploaded.

Usage
-----
There are two versions of the chunking API. Version 1 is the original version and version 2 was built as a backward compatible extension to support uploads directly to supporting target storages like S3. Version 2 is the recommended version to use.

Version 2 comes with a few additional requirements and limitations to consider (compared to version 1):

- Every request needs to have a ``Destination`` header present which specifies the target path of the file
- The naming of the individual chunks is limited to be a number between 1 and 10000
- The chunks will be assembled in the order of their names
- The size of chunks must be between 5MB and 5GB (except for the last chunk, which can be smaller)
- Chunks cannot be downloaded from the upload directory

Nextcloud will expire the upload directory after 24 hours of inactivity. This means that if you start an upload and do not finish it within 24 hours, the upload directory will be deleted and the upload will fail.

Chunked upload v2
-----------------

The API is only available for registered users of your instance. And uses the path:
``<server>/remote.php/dav/uploads/<userid>``. For this guide we will assume:
``https://server/remote.php/dav/uploads/roeland``



Starting a chunked upload
^^^^^^^^^^^^^^^^^^^^^^^^^

A chunked upload is handled in 1 folder. This is the location all the chunks
are uploaded to.

Start by creating a folder with a unique name. You can list the current available
folder but if you take a random UUID chances of collision are tiny.

.. code-block:: console
curl -X MKCOL -u roeland:pass \
https://server/remote.php/dav/uploads/roeland/myapp-e1663913-4423-4efe-a9cd-26e7beeca3c0 \
--header 'Destination: Destination https://server/remote.php/dav/files/roeland/dest/file.zip'
Uploading chunks
^^^^^^^^^^^^^^^^

Once a folder for the chunks has been created we can start uploading the chunks.

- The naming of the individual chunks is limited to be a number between 1 and 10000
- The chunks will be assembled in the order of their names
- The size of chunks must be between 5MB and 5GB (except for the last chunk, which can be smaller)


.. code-block:: console
curl -X PUT -u roeland:pass \
https://server/remote.php/dav/uploads/roeland/myapp-e1663913-4423-4efe-a9cd-26e7beeca3c0/00001 \
--data-binary @chunk1 \
--header 'Destination: Destination https://server/remote.php/dav/files/roeland/dest/file.zip'
curl -X PUT -u roeland:pass \
https://server/remote.php/dav/uploads/roeland/myapp-e1663913-4423-4efe-a9cd-26e7beeca3c0/00002 \
--data-binary @chunk2 \
--header 'Destination: Destination https://server/remote.php/dav/files/roeland/dest/file.zip'
This will upload 2 chunks of a file. The first chunk is 10MB in size and the second
chunk is 5MB in size.

Assembling the chunks
^^^^^^^^^^^^^^^^^^^^^

Assembling the chunk on the server is a matter of initiating a move from the client.

.. code-block:: console
curl -X MOVE -u roeland:pass
--header 'Destination:https://server/remote.php/dav/files/roeland/dest/file.zip' \
https://server/remote.php/dav/uploads/roeland/myapp-e1663913-4423-4efe-a9cd-26e7beeca3c0/.file
The server will now assemble the chunks and move the final file to the folder ``dest/file.zip``.

If a modification time should be set, you can by adding it as header with date as unixtime:

.. code-block:: console
curl -X MOVE -u roeland:pass
--header 'Destination: https://server/remote.php/dav/files/roeland/dest/file.zip' \
--header 'X-OC-Mtime: 1547545326' \
--header 'Destination: https://server/remote.php/dav/files/roeland/dest/file.zip' \
--header 'X-OC-Mtime: 1547545326' \
Otherwise the current upload date will be used as modification date.

The chunks and the upload folder will be deleted afterwards.

Aborting the upload
^^^^^^^^^^^^^^^^^^^

If the upload has to be aborted this is a simple matter or deleting the upload folder.

.. code-block::
curl -X DELETE -u roeland:pass \
https://server/remote.php/dav/uploads/roeland/myapp-e1663913-4423-4efe-a9cd-26e7beeca3c0/
Chunked upload v1
-----------------

The API is only available for registered users of your instance. And uses the path:
``<server>/remote.php/dav/uploads/<userid>``. For this guide we will assume:
``https://server/remote.php/dav/uploads/roeland``



Starting a chunked upload
^^^^^^^^^^^^^^^^^^^^^^^^^

Expand Down Expand Up @@ -60,7 +160,7 @@ Assembling the chunk on the server is a matter of initiating a move from the cli
The server will now assemble the chunks and move the final file to the folder ``dest/file.zip``.

If a modification time should be set, you can by adding it as header with date as unixtime:
``curl -X MOVE -u roeland:pass --header 'X-OC-Mtime:1547545326' --header 'Destination:https://server/remote.php/dav/files/roeland/dest/file.zip' https://server/remote.php/dav/uploads/roeland/myapp-e1663913-4423-4efe-a9cd-26e7beeca3c0/.file``"
``curl -X MOVE -u roeland:pass --header 'X-OC-Mtime:1547545326' --header 'Destination:https://server/remote.php/dav/files/roeland/dest/file.zip' https://server/remote.php/dav/uploads/roeland/myapp-e1663913-4423-4efe-a9cd-26e7beeca3c0/.file``"
Otherwise the current upload date will be used as modification date.

The chunks and the upload folder will be deleted afterwards.
Expand Down

0 comments on commit 6d3a740

Please sign in to comment.