[-] chrisprice@alien.top 1 points 10 months ago

There are a lot of factors. Manufacturing process, head, lubrication, insulation, material thickness.

Nobody can give you a specific answer without a forensic teardown. I would say it's probably the casing, refinements in head size, and noise canceling insulation inside the drive.

Drive noise measurement is always a frustrating thing to keep track of. It's much easier to just find a solution to isolate the drive noise completely and not worry about it.

[-] chrisprice@alien.top 1 points 10 months ago

There's no need to toss the drives at three years. Run drive diagnostics on them using a tool (GSmartControl, WinDLG, Hard Disk Sentinel, etc). Ideally every six months full scan, at least once per year.

Drives easily can last ten years without issue, and the odds of all drives failing simultaneously is near-zero.

Really you should keep at least one, ideally two, drives at different locations. And add an encrypted cloud backup to the mix.

[-] chrisprice@alien.top 2 points 10 months ago

Encryption doesn't have a major drive impact. The data is effectively the same.

Now, if you set a drive to maximum encryption, and encrypt all sectors - that's basically going to force the drive to write to every sector. This will uncover any drive surface errors, and it's basically the most stress test-y thing you can do to a drive. If there are bad sectors, you can bet pending reallocated sectors will go up.

Again, that does not mean the drive has failed. Those bad sectors could have been there since the factory. 75 is a concern. But is not a failure.

The fail alert is that you need to zero/erase all sectors, which will allow the drive to do a reallocation. When an erase bit is sent to the sector (by the OS/erase command), that's when it will reallocate.

[-] chrisprice@alien.top 2 points 10 months ago

This is why we tell people to have one working copy backed up to two separate backup drives (ideally in two places), and a cloud backup too.

Easily can have two drives fail.

Now, if it happened all the time, manufacturer warranties would be 3-6 months and drive makers would say "you'd better buy ten and have nine backups!"

Because... the alternative is to print everything out on paper.

[-] chrisprice@alien.top 1 points 10 months ago

BD-RE XL you mean. Sorta. It was/is much more like DVD-RW functionally than DVD-RAM.

DVD-RAM was truly trackless MO. You could format one with any HDD/SSD file system.

BD XL is the best consumer archival media out there. But it's not cheap compared to stashing a few 18TB drives in different locations with a Faraday cage.

[-] chrisprice@alien.top 1 points 10 months ago

The good news is HAMR drops next year, and that will substantially lower prices.

WD has all but said they will not be able to answer HAMR next year, and will have to compete on cost with smaller drives.

And Toshiba and HGST have been silent.

All that adds up to Seagate getting to 32-40TB next year, and the rest of the players having to slash prices so buyers can get two drives and save a lot.

[-] chrisprice@alien.top 1 points 10 months ago

If you backup /obb's, it usually won't. From Android 1.0 - 11, that should be fine.

The hard part is with newer Androids that block access to the /Android/data/obb trees.

[-] chrisprice@alien.top 1 points 10 months ago

Off site backup for your existing data, in case of disaster.

Sorry to burn $200, but you'll thank me later when disaster strikes... living in an area where five neighboring cities all burnt to the ground in the course of 24 months (in the three worst wildfires in California history, back-to-back-to-back)... I have experience here.

[-] chrisprice@alien.top 1 points 10 months ago

Cleared out my store's three units.

Pairing this with 8TB SATA SSD for disaster proof backups. After twenty years, getting all my data in one place...

... Well, except the 20TB overflow. That backs up to my $279 20TB's. Which hook into BackBlaze. They had better appreciate the loyalty. The 28TB backups sure cost them a lot.

I'm just glad I don't have more. Then I'd have to set up a NAS and, ugh, I'm just working too hard to enjoy doing that these days.

Only hiccup so far is either my 8TB Samsung 870 QVO died or the enclosure, a whopping 30 minutes in. This is why I backup meticulously in (at least) three separate locales, plus encrypted cloud.

[-] chrisprice@alien.top 1 points 10 months ago

There are excellent articles that go over all this. Do a something search.

Bottom line, yes, you should at least do 3-2-1 methodology. More than that is gravy.

[-] chrisprice@alien.top 1 points 10 months ago

I still hope to one day make a desktop OS do full system restore that uses rclone. Break system, buy new one, feed decypt key and server info during OOBE, click go. Desktop restored. Completely.

We can do it. We literally have the technology.

[-] chrisprice@alien.top 1 points 10 months ago

Legally, if they are hosting compromising things of any sort and don't report you for it they are in big trouble. So expect that.

That's not true. What keeps Mega from being shutdown (like MegaUpload was previously), is that Mega is carefully following Australian and American laws - which do safe harbor cloud providers that host fully encrypted files.

Now, if Mega receives a copyright infringement report that includes the decryption key... then they are obligated to investigate. This is why pirated files hosted on Mega with the keys posted pubicly, are so often taken down.

It's not that Mega is decrypting the files on the backend, it's that content providers are searching for the keys and sending them to Mega when they find them.

Apple considered decrypting iCloud Photos, despite no legal obligation to do so, because of political pressure. They backed down when consumer/EFF pressure changed the narrative.

view more: next ›

chrisprice

joined 11 months ago