Since moving to an Enterprise server(s) for database/web we have still since a form of image upload corruption, and we think we know why. Maybe. We have heard from all walks of Wordpress user about this, and though our new cloud host has done an amazing job of speeding up our set and custom configuring all settings, there is probably a simple and fundamental flaw in Wordpress.
We were seeing instances where the post had been published, and images weren't done being resized on the server, and this is a quite powerful server. You could reload the page and the image would then have finished being re-sized. Sometimes it just gets stuck not having been resized, possibly due to W3TC caching? but we have seen this with just one image as well. Some diagnostics must be involved in this project. We have a dev. server setup which is identical to our production run.
I think it has to do with upload bandwidth personally because this would explain that the server is only going as fast as the data you provide it...posting the actual ad is almost instant.
The process now is like this: publish post, upload images then resize.
My current dev. suggests this:
"WP or CP? .....they publish the post first because you need to have the post ID, once you do you can link the images to that post, otherwise you just cant
what I think should be done is to upload the images resied already, instead of uploading at the original size and then resize them
I think thats the way to go, because posting images and then publishing the ad is a dead end"
While I am a luddite, it would seem to mean if we could have the images resized first on the users' PCs, then uploaded, the web server load would lightened and a disaster would be averted.
By that I mean, if we are seeing this with one user posting at a time.....it could get ugly when things scale up.