Comments on: Backblaze Drive Stats for Q1 2023

By: Meda

Meda — Sun, 25 Jun 2023 18:36:20 +0000

What about the WD Black 2TB ? is it reliable enough ?

By: Carlos

Carlos — Mon, 22 May 2023 11:12:16 +0000

Are you also gonna do Q1 SSDs report, or is that a yearly publication or something like that?

By: Robert

Robert — Sat, 13 May 2023 11:57:48 +0000

But brief look at difference in Seagate vs WD at 16T shows that there is strong reliability towards WD

By: Robert

Robert — Sat, 13 May 2023 11:55:28 +0000

It would be helpful if there would be chart which shows failure rate of manufacturers per volume. It would show simillar disk that happened to be produced and possibly used at same time

By: jdrch

jdrch — Mon, 08 May 2023 22:29:27 +0000

How does BB define “failure?” For example, I’ve had a 12 TB Seagate Barracuda Pro in CrystalDiskInfo “Caution” status for approaching 3 years now, with no symptoms. None of my other large capacity (10+ TB) HDDs (Seagate Exos, Toshiba MG, WD Gold, WD UltraStar) have shown as much as a hiccup during that time.

To me, this data isn’t as indicative of HDD reliability as it is that modern datacenter workloads are beginning to eat significantly into HDD life.

FWIW, assuming the HDDs used are datacenter drives, they’d be under warranty at the time of failure anyway.

By: Francisco Comparatore

Francisco Comparatore — Fri, 05 May 2023 10:30:25 +0000

Thank you very much. It was a very interesting reading

By: Kevin

Kevin — Fri, 05 May 2023 04:46:54 +0000

In reply to Granny.

agreed – I mean the annual failure rates at a data center aren’t going to be that relevant to a home user right? not if drive failure has anything to do with usage.

So maybe rather than time, it would be more useful to score drive failure against MB or TB written and read.

By: Granny

Granny — Thu, 04 May 2023 20:35:24 +0000

In order to give a better picture, you should add some details about drives, along with power on hours:
– workload (egress + ingress);
– temperature conditions (avg external temp, avg internal temp, min and max temps);
– room humidity;
– more details about failures: type of failure, hardware part that failed (head, controller, motor, platter,…) etc.

By: John Haller

John Haller — Thu, 04 May 2023 15:55:15 +0000

In reply to Andy Klein. Given the annual failure rates, but without looking at the dataset, it appears a good majority of the drives made it to retirement without failing. The failure rate and cost of operating the drive (power, real estate, etc) as opposed to a bigger drive presumably both played into retirement decisions for particular drive types. I've got 10 drives in one system with two clumps at 7 (6TB) and 9 (4 TB) years old. I'm guessing it's about time for retirement, and will be able to drop the number of drives to about 4. I just hope I don't get correlated failures.