Comments on: Backblaze Drive Stats for Q3 2022

By: Hok

Hok — Thu, 23 Mar 2023 03:10:54 +0000

Very interesting data…but when you refer to disk failures what do you mean? that disk stops working? it is removed? and then how do you know that they have had those errors? because they could not write or read? I know…all of a sudden I have asked a lot of questions! It is simple curiosity, to know how you know. I suppose that no data is lost when these failures occur in the disks, I have my data in your disks!! .

By: Andy Klein

Andy Klein — Mon, 30 Jan 2023 12:35:30 +0000

In reply to Dimitar Ivnov. Most drives are rated to 60° C, but that's for a short duration. If the 52 was for a prolonged period, days multiple days, that's not good. 43-45 is fine, although again over a long period (weeks) it could be an issue. In general, we try to keep drives below 40 on average.

By: Dimitar Ivnov

Dimitar Ivnov — Sat, 28 Jan 2023 15:40:32 +0000

Hi guys! Is it normal for ST8000NM000A to reach 52°C = 125.6 F in external usb encloser?
In the PC now reach 43-45 °C.

By: Andy Klein

Andy Klein — Wed, 18 Jan 2023 13:43:18 +0000

In reply to takeos. We buy the most cost effective drive available to us. The 18 and 20 TB drives are still more expensive is their total cost versus the 14TB and 16TB drives.

By: takeos

takeos — Tue, 17 Jan 2023 07:13:44 +0000

Is there a reason why there are no 18+ TB drives (would have been available for more than a year)? Are you waiting for something “better”?

By: Ralf

Ralf — Thu, 01 Dec 2022 05:41:50 +0000

While happy with my hgst 4TB at the moment I started lookin’ for the HUH728080ALE600 and
HUH728080ALE604 on ebay.

Despite their great productionnumber there are only few to be found. Has there been some renaming of the drives due to WDs acquisition of HGST? Or after they had to sell part of their production to Toshiba?

Anybody know the new designations?

By: Ralf

Ralf — Thu, 01 Dec 2022 05:36:07 +0000

In reply to Dinith Dulanjana. Bought those buggers from a data center refurbisher with alreade 40' hrs on but they are running like a charm for 2 yrs now. Absolutely great product

By: Hank Barta

Hank Barta — Sun, 06 Nov 2022 10:52:35 +0000

Thank you so much for continuing to publish this information. To put this in context, can you characterize the typical service these drives see? My expectation is that the drives are written infrequently and then occasionally read as backups are restored. Of course churn in your customer base would result in deletion and rewriting. Likewise with updates to individual backups, but even this case seems like it would involve only a small amount of data compared to a client’s total backup storage. Nevertheless, I expect that the duty seen by these drives would be lighter than, for example, a system that hosts a database with near constant updates and additions.

Do you think that usage patterns affect drive service life? If this is something you can discuss, does Backblaze Engineering do anything to manage usage patterns to prolong service life?

Thanks!

By: Disqus User

Disqus User — Sun, 06 Nov 2022 00:01:54 +0000

Andy, you guys are doing the industry an amazing service by openly sharing this information. One question that I’ve wanted to understand – does more disk usage increase failure rates? Any difference between reads and writes? Even head movement – would a “busy” head with lots of seek activity cause more failures?