Discussion Forum

tsamant

Member

From: Chicago

Registered: May 14, 2015

Posts: 52

A file browser problem

We have a file browser in admin that is backed by S3.

An admin just managed to "delete" a folder. As far as I can see, this means that some rows in bigtree_resource_folders and bigtree_resources have disappeared.

Thanks to S3, the site continues to serve files that were in the "deleted" folder. But they are now unavailable from the file browser.

I can restore the rows that were dropped from the two tables. But is that the right thing to do?

Best,
Tushar

timbuckingham

Administrator

From: Baltimore, MD

Registered: April 2, 2012

Posts: 974

Re: A file browser problem

If those files are still serving correctly from S3, restoring those dropped rows is the correct thing to do.
Looking at the code, it should have also deleted the files from Amazon S3 -- so there's a bug to check on as well!

tsamant

Member

From: Chicago

Registered: May 14, 2015

Posts: 52

Re: A file browser problem

Yes, there is indeed a bug; we have seen that upload-with-replace to S3 does not replace, but creates a file with a fresh numerical suffix.

No complaints this one time! But we do have a growing number of unnecessary files.

timbuckingham

Administrator

From: Baltimore, MD

Registered: April 2, 2012

Posts: 974

Re: A file browser problem

I wasn't able to replicate the numerical suffix issue unfortunately. I did fix the deletion of resources failing, though!
https://github.com/bigtreecms/BigTree-C … 16dea3db33

This was due to a change to allow for protocol agnostic // URLs for cloud resources. The deleteFile call wasn't changed to understand them so it assumed they were local files.