If you take photos or create media for clients, what are you using for a long-term storage solution? Even though I’m using a cloud solution now, I’m looking to get away from them and keep everything local.
Backstory: I’m currently using an Amazon S3 Bucket. I have a management rule in place that moves current to S3 Glacier after 30 days. I’ve had the system in place for a number of years. It’s been easy to upload content to. Recently, I’ve need to pull stuff out of the S3 and I’ve been running into an issue – restoring objects at scale.
Francis Ford Coppola (The Godfather, Apocalypse Now, etc) lost 15 years of work and family photographs when thieves stole all his computers and his backup drive in 2007.
Thankfully I’ve only restored in a testing capacity; it was uneventful/simple. They also have options to mail a flash drive $99/~1GB or a USB drive $200/4TB for larger/full downloads.
How are you moving data to/from Glacier? I’ve done several restores in the multiple TB range without issue. It’s hard to beat the price of Glacier ($1/TB/Month). I think I’ve got north of 60TB stored there for my day job.
I work in post for corporate media, and we do a variation of all these and then we are looking to start offloading to LTO in the next year or so.
The thinking is to treat our NAS’s like “time machines” not in an Apple sense but in they only need to store 3-5 years of our most recent work, and some archival exports of older projects. Everything else will get offloaded to LTO.
That’s an excellent question. I upload to S3 I use “S3CMD” on the command line for use Forklift. I have a management rule in place that moves anything that hasn’t been touched in 30 days from S3 Standard Infrequent Access to S3 Glacier Flexible Retrieval.
For the restore, I’m using the S3 console. I’ll admit my problem is mostly user error. I wanted to go in and just select the objects I wanted to restore in the same matter as restoring a single object.
I’m going down the batch restore route. I finally got an Inventory to run on the entire bucket and a second one to run on a prefix that I’m testing. I think I finally got the IAM Role and Trust Policy figured out. I’ll see when my test batch restore runs.
If you have any tips, tricks, or recommendations, please let me know.
To go full circle on this. My latest restore attempt was successful. I think I’m okay with my current S3 solution for cloud back-ups. Though this episode has made me realize I need an additional backup that’s more immediate.
TL;DR: Thank you for your suggestions here. I will explore some of them.