Should I disable proxy for my 150k imgur images?

[Mave]

I've been proxying images for the past 3 years.
However I have over 150k imgur images embedded on the forum. The image_cache folder is also really large because of this.

Should I disable the proxying of images for all imgur urls?

My backups now they an insane amount of time.
Follow up question: OR should I just never backup internal_data?

Thanks.

 

[TPerry]

You want to be able to restore your site fully? You BETTER back up internal_data.
It's simply one of those necessary directories. Hell, if you want to merge another site to yours, there are two directories you better have access to... internal_data and data.
If the directories are getting to large too back up because of embedded imagesvia proxy then you can either disable the proxy and if the remote hosted site goes down or changes the image it no longer shows (or possibly shows a different image if updated, a way of sneaking porn/advertising into some sites).... or require all images be uploaded directly (and here you are back to the initial issue with back up sizes).
Simple fact of life is... if you allow images (and many of them) on your site, you are going to need storage.
All the proxying of the image does is make it a "little" faster for display since it is served locally.. you are STILL at the mercy of what is being served on the remote end once your cache period expires and it retrieves a new copy.

 

[digitalpoint]

You would want to backup at least some of internal_data. You would want to backup internal_data/attachments unless you are okay with losing attachments. The rest isn't so important. Things like image_cache is just that... a cache. It's not meant to be permanent, so losing it because you didn't back it up isn't catastrophic. It just means the image proxy would refetch them if you lost them.

Another thing you might want to consider is rather than doing a full backup every time (if that's what you are doing), is use a backup system that supports diff (only need to backup new files then). As an example, I use rsync for my backup of files to an offsite location. Backs up things like new attachments, deletes from backup the files that have been deleted without needing to do a full backup every time (I also use multiple sets so I can go back to a set that included a file that might have been inadvertently deleted, and then deleted from the latest backup). Doing a full backup every time will just make your backups take longer and longer and longer as time goes on with things like attachments and avatars always growing.

 

[TPerry]

nd then deleted from the latest backup). Doing a full backup every time will just make your backups take longer and longer and longer as time goes on with things like attachments and avatars always growing.

unluckily, most of those doing backups are not to your or my level.
They simply back up the entire structure because that's what the "point & click" has set up.
At one time I did 'diffs for 6 days and on day 7 did a fresh backup. I found that honestly doing (for my smaller site) a full backup was not that much more intensive and took less steps to restore from if I need to later.