theodi / shared

Repo that we use for non-repo-specific stories and other shared stuff.
22 stars 1 forks source link

Handle large CSV file sizes for upload #710

Closed Floppy closed 9 years ago

Floppy commented 9 years ago

We can't validate anything over 700k in a file upload without causing an error (see https://github.com/theodi/csvlint/issues/201). It's unclear whether this is because of the nginx file upload size limit, or something in the app instead.

We want:

Floppy commented 9 years ago

@bcouston and @quadrophobiac are going to take a look at this to see if we can make some progress before CSVlint is used for the DaPaaS workshops next week, or at least isolate the problem further.

bcouston commented 9 years ago

@Floppy This might be server side, locally I can validate a 1.2MB csv file (http://training.theodi.org/resources/dataset2.csv), on live, 413 entity too large error with an infinite validation loop.

bcouston commented 9 years ago

On the local side, Mongo can't store files larger than 16MB, investigating into using a package that splits data into chunks (http://www.resumablejs.com/) or GridFS.

bcouston commented 9 years ago

The error

screen shot 2015-08-21 at 14 09 52

dicknorris commented 9 years ago

Book to DaPaaS > WP1 > Task 1.2

Floppy commented 9 years ago

The server-side and investigative stuff here is OK, and the next thing is to look at how we make a better upload experience for people using a javascript uploader like ResumableJS.