Interesting how the numbers carry over year-to-year in<p><a href="https://www.backblaze.com/blog/wp-content/uploads/2020/02/Blog_3_year_Drive_Stats_Chart.png" rel="nofollow">https://www.backblaze.com/blog/wp-content/uploads/2020/02/Bl...</a><p>Some models are dwindling. Some are being tested. Others (like the Seagate and HGST 12 TB) are increasing. Only thing that's really perplexing is why they keep buying more and more of the high-failure-rate Seagate 12 TB drives. It must be more than 3% cheaper to buy (and service!) a Seagate with a 3% chance of failure than to buy an equivalent HGST with a 0.4% chance of failure. I guess when you have 120,000 drives, easy hot-swap enclosures, and software to handle it all that makes good sense! But as an individual consumer, even with a Backblaze backup, it's definitely worth my time to spend a bit more on a drive that's far more reliable than to save a few dollars on a Seagate.
Why do people use Amazon S3 when Backblaze B2 is 1/4 the cost of S3 and also includes a CDN for free. You also get way faster access speeds with Backblaze vs Amazon since they tier their IO speeds.<p><a href="https://www.backblaze.com/b2/cloud-storage.html" rel="nofollow">https://www.backblaze.com/b2/cloud-storage.html</a>
Does anyone remember what is their definition of "drive failure"? Is it SMART "failure imminent" report, single uncorrectable read error or complete data loss for a whole disk? I recall reading about it in one of their previous report, but can't find it again.<p>EDIT: nevermind, found it.<p>"Backblaze counts a drive as failed when it is removed from a Storage Pod and replaced because it has 1) totally stopped working, or 2) because it has shown evidence of failing soon.<p>A drive is considered to have stopped working when the drive appears physically dead (e.g. won’t power up), doesn’t respond to console commands or the RAID system tells us that the drive can’t be read or written."<p><a href="https://www.backblaze.com/blog/hard-drive-smart-stats/" rel="nofollow">https://www.backblaze.com/blog/hard-drive-smart-stats/</a>
Ah yes, the reliable BackBlaze folks. That they've out-Googled Google in a niche using mostly commodity infrastructure and kept their business alive for so long is a testament to their ingenuity (I wonder how their operating costs compare with AWS Glacier which has a theoretical advantage of unpowered disks.). And the releasing of this proprietary operational business data is a testament to their coolness factor.<p>It's a timely article as I'm looking at HC530's (WUH721414ALE6L4 / WUH721414ALN6L4 (wiredzone carries it)) for a home FreeNAS box:<p>- any relatively-modern enterprise 4U 3.5" storage box with Xeon 4 cores or so<p>- quieter, high-volume fan mod<p>- RAM: 64-128 GiB, beyond that isn't useful unless deduping<p>- NIC: X710-T4L 4x 10GbE copper NIC<p>- ZIL: mirrored pair of high-endurance, write-intensive, reliable SSD like Optane 900p/905p 280-480GB<p>- L2ARC: striped pair of read-intensive/larger SSDs like the Gigabyte Aorus Gen4 1 TB<p>This will fit nicely as my home NAS for a water-cooled dual EPYC virtualized server/workstation build underway. I managed to get a single water block with (3) G1/4 connections that will cool both CPUs and the VRM chokes/converters.<p>If anyone has better suggestions, please chime in.
Slightly off topic: is anyone using B2 (which seems cheaper if you have more than one computer for a certain amount of data) for personal data backups with strong client side encryption across multiple platforms (Linux, Mac, Windows)? If so, how do you handle it?
Looking at those Data,<p>It seems they will soon reach 1000 PB / 1EB.<p>The top 5 Annualised hard drive failure rate are all from Seagate. All Drive from Hitachi and Toshiba has AFR lowered than 1%.<p>So basically dont buy Seagate.
Does anyone here have experiences with BackBlaze's B2 service for hosting files? I'm considering switching to it from S3 because it is much cheaper. (I need to transfer 2-3TB / month, usually in 2-3 bursts of worldwide distribution).
I have made all my hard drive purchasing decisions based almost entirely on these reports for the last couple years and have not been disappointed with the results.
No really related remarks about this handy study, but anybody else still in real awe about how spoiled we are with regards to the sizes and speeds of HDs nowadays? I mean the smallest capacity drive on their chart is 4 <i>Tera</i>bytes.
I have about 10 TB of video files. I use BackBlaze for Windows but I would like the files to be available on other computers and my phone in my local network.<p>What can I use to do this and still keep offsite backups?
I have two ST12000VN0007 (VN) Seagate drives. The report shows the ST12000NM0007 (NM) has a 3.32% failure rate. I wonder how closely related the VN and NM models are.
Seagate always seems to have much higher faliure rates compared to HGST/WD/Toshiba etc<p>Does anyone here know the exact reason why? I assume there are enough people on this site who have worked for them or a competitor :)
Does anyone have any opinions and experience using backblaze as a personal only cloud storage and offsite backup for smaller amounts of data (under 30 TB)
I have wondered about system downtime or time operating in a degraded state.<p>My understanding is other than mirrored, RAID configurations may take a long time to rebuild on the larger drives and this is a contributing factor to why the highest sales volume of drives has been 'stuck' at 4TB (thus the lower $/GB price).