4 ms·
My recipe is "Go to Costco, buy the biggest backup drive they have, shuck the drive and put it in the NAS. When one dies, repeat." I end up with a great varie
by jedberg 6y ago
My recipe is "Go to Costco, buy the biggest backup drive they have, shuck the drive and put it in the NAS. When one dies, repeat."
I end up with a great variety of drives and firmwares that way.
I'm still waiting on that 2TB drive to die so I can put in a bigger drive and expand the array!
- ed25519FUUU 6y agoI’ve heard that many external drives can’t be shucked any longer, as they don’t include proper sata connectors. Have you encountered that?
- jedberg 6y agoI have not yet run into that. So far they've all been fine. The shucking has gotten harder for sure, I need a lot more force to pry them apart now. So much that I worry about damaging the drive.
- whoopdedo 6y agoSomeone knows the trick. Fun anecdote: I bought a 8TB Seagate from a local retailer at discount because it was an open-box return. It looked fine until I plugged it in and the drive showed up without the usual setup.exe and the drive was partitioned oddly. Okay, go to resize the drive and only see 500GB of space. Defective drive? Maybe that's why it was returned. Then I pull up the SMART details. The drive reported over 10,000 hours on it, a history of read errors, and was manufactured by Toshiba. I looked over every millimeter of that exterior and couldn't see the slightest mark of a pry tool or even a sticker being out of place. I'm impressed if they were able to open the case and reseal it that cleanly. Although I also theorize that someone is stealing empty cases from the factory that they assemble with "recycled" drives. Either way, it was a professional crime. Not like the lazy fraudsters who just put a literal brick in the box and hope the store employee doesn't look inside. End of story, I got my money back without problem since the store manager knows me.
- exikyut 6y agoDid you tell them what you'd discovered? (If yes, what did they think of the situation? (Did you get any sense where the drive came from?)) Idle thought experiment: maybe the enclosures were cloned. (ie, the originals were 3D scanned + cleaned up + etc.) I wonder how much that would cost end-to-end.
- rangibaby 6y ago> Idle thought experiment: maybe the enclosures were cloned. (ie, the originals were 3D scanned + cleaned up + etc.) I wonder how much that would cost end-to-end. Do you think it's worth the effort of 3d printing, painting, assembling everything, etc just to swap your fake for one real HDD?
- whoopdedo 6y agoLike I said, I'm friends with the manager and he was the one who told me about nearby stores being hit with a box-of-bricks return. Looking at the authentic model I eventually got, it's in three parts with the outer shell a glossy black and and insert that's slate colored with vent panels for passive cooling. And a Seagate logo inset that doubles as an power/activity light. The bottom has small rubber feet and the sticker with model and serial number on it. They didn't have another drive in stock to do a side-by-side comparison with, but I couldn't notice anything that would have tipped me off it was a fake. Even tried prying at the seams to see if it came apart any easier and it was held just as firmly as the real thing.
- ValentineC 6y agoIn my experience, only Western Digital portables (2.5") have direct USB connectors.
- jolmg 6y agoI did that once, and found that even when connected directly via SATA, the drive seemed limited to USB3 speeds... I guess that's cost reduction? Like, why make a faster drive when it's meant to be accessed only via USB3? It was one of those Seagate Slim 2TB external drives.
- bufferoverflow 6y agoUSB 3.0 speed limit is 5Gbps. That's 625 MB/s, which is faster than any HDD I'm aware of. Modern USB 3.2 can do 20 Gbps. That's 2500 MB/s.
- jolmg 6y agoI'll check what the speed was again, when I get back home.
- jolmg 6y agoOk, so I got it connected via SATA. I tried the following: pv /dev/zero > /mnt/foo umount /mnt # trying to get rid of any caching of foo. mount /dev/sda4 /mnt pv /mnt/foo > /dev/null pv /dev/sda3 > /dev/null pv /dev/zero > /dev/sda3 All pv invocations showed an average of ~126MiB/s each. I don't have another HDD on hand to compare since I've upgraded everything to SSDs. Though I don't remember what's normal for HDDs, to be honest 126MiB/s doesn't seem that slow. I thought I remembered something lower.
- fulafel 6y agoNot relevant here but a tangent: In USB 2.0 times the efficiency was pretty terrible, you could get ~25 MB/s disk transfer throughput from the 480 Mbit/s signaling rate. It improved a little over the years... I wonder what are the actual speeds you can get from USB 3.0/3.2.
- bufferoverflow 6y agoLook at USB 3.2 SSDs and NVMe enclosures. They go over 1000 MB/s for sequential reads. https://www.anandtech.com/show/14661/usb32-g2-portable-ssds-roundup/3 https://www.anandtech.com/show/14661/usb32-g2-portable-ssds-... I'm pretty sure they are limited by the USB <--> NVMe controller, not USB itself.
- bufferoverflow 6y agoCostco currently has an 8TB drive for $140, while Amazon has one for $135 with free delivery. And it's a Seagate drive, they don't do so well in Backblaze reports.
- jedberg 6y agoYeah they are always loser drives in the back blaze report, but so far it's been fine. No more than an expected level of loss.
- boring_twenties 6y agoI spoiled myself with a 10TB for $150 from Best Buy last year. Now I need another one but I can't stomach a worse deal than that!