-
Type: Improvement
-
Status: Resolved
-
Priority: Minor
-
Resolution: Fixed
-
Affects Version/s: 4.1.4
-
Fix Version/s: 4.5.0
-
Component/s: Remote client
-
Epic Link:
-
Tags:
-
Sprint:nxDrive 11.2.5
-
Story Points:1
Current Behavior
An upload is done in 3 steps:
- sending chunks
- completing the upload on the server
- calling FileManager.Import to effectivly create the document and attach the blob
When the step 1 is done, it cannot be done again: the server will send the uploaded chunks list.
When the step 2 is done, it can be done again if the step 2 or 3 failed. If it failed at step 2, it is legit to retry. but if it failed at step 3, retrying to complete the upload will end on error, or worse, generate more work on the third-party provider or Nuxeo itself.
When step 3 is done, it cannot be done again as the upload is then removed from the database. But if it failed, it will be retried.
The Issue and Fix
If the step 3 failed, then step 2 will be done again. And this is not what we want. Instead, we should store the upload datus in the database to let the application knows that it should go to the step 3 directly.
- causes
-
NXDRIVE-2412 Clicking on pause and resume make it restart from the beginning
- Resolved
- is related to
-
NXDRIVE-2298 Improve large file upload resiliency
- Resolved
- Is referenced in