tenchiken

joined 2 years ago
MODERATOR OF
[–] [email protected] 1 points 3 hours ago

Different groups selling different things. OpenStack still around, albeit a shell of it's former scale

[–] [email protected] 2 points 3 hours ago (1 children)

.... You do realize that they still have hundreds of thousands of VMs in their OpenStack services? Those are VMs too.

Hell back in 2008 Slicehost had more than 40k VMs before Rackspace bought em.

Wait till you hear about places like AWS or Azure....

[–] [email protected] 1 points 4 hours ago

Unfortunately the client seems to explicitly skip network shares, so unless there's some trickery I don't think that is viable.

Seems explicitly against TOS too now:

https://www.backblaze.com/docs/cloud-storage-network-attached-storage-devices

I'm all for any workaround, but worth noting so we don't get OOP caught out unexpectedly

[–] [email protected] 3 points 10 hours ago

This! Tape is still the golden standard for high capacity!

[–] [email protected] 2 points 10 hours ago

I use b2 for about 15tb, still one of the cheapest really without being sketchy. Cost isn't too bad unless you are reading it often.

As another person already noted, if you really need to back up high amounts, tape is the way to go. Plan to keep your critical stuff off site somehow too. For large amounts sneaker net is still best unfortunately.

[–] [email protected] 2 points 12 hours ago

I've never had issues; maybe been lucky lol.

That said, they provide some amazingly detailed status about drives! Worth looking at the reports they post. Might get you insight into what to expect from the various manufacturers and models... Maybe avoid some junk drives in the process.

One of the most recent of these:

https://www.backblaze.com/blog/backblaze-drive-stats-for-q3-2024/

Raw statistics might help cut through a lot of bias!

[–] [email protected] 3 points 14 hours ago

Oh important note! Check to make sure any drives you'll use are CMR, not SMR (shingled)! SMR will not function right in raid and will fail from arrays.

[–] [email protected] 15 points 14 hours ago (11 children)

I'd start by noting that raid is more about availability, not backup... I suspect you already have that in mind but just in case. Ideally if you are up for learning ZFS, that is one of the most resilient raid tools out there. Most NAS and Unix or Linux OS will have support for this.

Never connect RAID disks via USB... This only causes headaches.

Avoid SATA port multipliers, these can cause problems in raid.

SAS has the most reliable and flexible options for connectivity. Used JBOD chassis, even small, can be found cheaply and will run SATA disks well.

As to cloud data, I strongly recommend BackBlaze. Many utilities can natively interact with it (API compatible with Amazon s3) and you can handle encryption on the fly with several sync options. They are one of the cheapest solutions, and storage is pretty much all they do.

With pretty much any cloud storage, look at the ingress/egress cost of your data too... That is where many can bite you unexpectedly.

Worth noting that when you get to large storage, a good organization method for your data is key so you can prune and prioritize without getting overwhelmed later... Don't want several copies of the same thing eating cash needlessly.

Good luck! And welcome to the wonderful illness known as data hoarding!

[–] [email protected] 39 points 19 hours ago (5 children)
[–] [email protected] 15 points 1 day ago

I'm not trans, but I can say that no matter the various reasons, my personal one was simple:

Everytime someone replied to one of my comments with the most asinine, absurd or just downright stupid asshole approaches, it was always hexbear.

Literally a sigh of relief when Lemmy added personal instance blocking. Immediately hit that button and never looked back.

[–] [email protected] 51 points 1 day ago* (last edited 1 day ago) (7 children)

For what it's worth, your instance name is spectacular!

view more: next ›