in nfs mount, upload gets restart if upload data makes storage of nfs server exhaustive

[bishwasojha]

Describe the bug

In nfs mount, the uploading does not get completed if the uploaded data makes storage of nfs server exhaustive. During uploading, the data is synced to the nfs server, and the server storage is occupied; however, once the server storage is exhausted, the progress bar of uploading restarts from the beginning, and the server storage that was previously occupied is no longer occupied.

Steps to reproduce

Steps to reproduce the behavior:

1. setup nfs client server https://owncloud.dev/ocis/storage-backends/dcfsnfs/
2. upload data exhausting the storage of the nfs server

Expected behavior

• uploading terminates with some notification

Actual behavior

• The progress bar during uploading goes to completion, but then the progress bar restarts from the beginning .

Setup

• setup nfs client and nfs server

OCIS_VERSION=ocis:2.0.0-rc.1

Additional context

Screencast.from.14-11-22.12.27.52.+0545.webm

[bishwasojha]

• The uploading restarts with status 507 and 404 Status

Screencast.from.14-11-22.04.29.57.+0545.webm

[micbar]

@butonic @kulmann Can you qualify if this needs cross platform changes?

[butonic]

AFAIU this is the web ui retrying the POST / PATCH request during tus uploads.

On the server side, when we run into en error during a chunk upload we purge the upload. This happens in the PATCH request when it returns 507.

That is also the reason why the subsequest HEAD request receives a 404 ... the upload really is gone.

The desktop client retries the upload with a backoff ... I don't think the web ui should retry on a 507, should it?

[JammingBen]

I don't think the web ui should retry on a 507, should it?

If the upload in gone on the server side, we should avoid retrying, yes. It should be pretty easy, we just need to hook onto the onShouldRetry event and check for the status code 507 (https://uppy.io/docs/tus/#onShouldRetry-err-retryAttempt-options-next-gt-next-err).

@butonic Are there any more status codes where retrying should not happen?

[butonic]

Hm, InsufficientStorage 507, RFC 4918, 11.5 ... actually any 5xx?

[JammingBen]

#7985 has been merged on the stable-6.0 branch (coming to master soon as well) which addresses this.

Edit: Now in master, hence I'm closing this issue. Feel free to re-open if this is still an issue.