![]() Their goal is to store as much data as cheaply as possible, and SAS SSDs are the antithesis of that. That's very different from skimming drive/SMART data and seeing that a drive is starting to fail and being proactive about replacing it.Ĭlick to expand.Remember that Backblaze's priority is their business of storing data - compiling drive statistics is just something they do as part of that. Outright controller failures are less common now, but they're terrifying zero-warning events that are generally unrecoverable. Considering they also have a bunch of BOSS arrays in the data, that sounds like they're boot drives or have some other high speed, high reliability component to the use of at least some of the drives.Īlso since I normally follow the HDD stats more closely, I'm kind of annoyed that they don't separate out the two kinds of SSD failures- The drive's endurance is exceeded and the drive is out of spare sectors/has gone into read-only mode from old age vs. If you really need speed, that's a reasonable compromise over the highest density HDDs. The <1TB capacities aren't honestly that expensive and you can go pretty big- ~6-7TB for under 2 grand. ![]() I know they're all about punishing cheap commodity drives, but I'm kind of surprised they didn't use say some 12 Gbit SAS SSDs. ![]() That's anecdotal but would be really helpful and interesting to know. Exactly how long did it live? Was it idle for some of the time? Did it die by throwing errors or did it just have a controller failure and poof dead. The presentation on that one is really confusing. I know that's not how statistics works but that poor single 250 GB class Seagate that crapped the bed after 4 months looks really bad in their data. $/GB is important, but they have to be getting some sweetheart deals on the ~250 GB class for that to make a lot of financial sense. In 2020, the failure rate was 0.93% and in 2021 it was 1.10%.I'm honestly shocked they're not just using commodity drives, they're using what are generally among the lowest end commodity drives. This increase is due to the age of the drives, as all hardware becomes more likely to fail over time. The yearly failure rate for hard drives has risen in the past three years, reaching 1.37% in 2022. Additionally, these drives had only been in use for a little over 22,000 days and were mainly used as backup drives to replace failed drives. It had no failures recorded in 2022, but it's important to note that there were only 79 of these drives in use. Seagate's 8TB hard drive, the Exos ST8000NM000A, was the standout performer in the report. The HGST Ultrastar He8 8TB and Seagate Exos X14 14TB had the highest failure rates at 5.27% and 5.70%, respectively, but it's important to consider the smaller sample size for these models. This is due to the aging of the drives, which become more prone to failure over time. The annualized failure rate for hard drives has increased in recent years, with the rate for 2022 reaching 1.37%. However, it's important to note that Backblaze only had 79 of these drives in use. The company analyzed 230,921 hard drives and found that Seagate's Exos 8TB (ST8000NM000A) had no failures in 2022. The report provides information on the reliability of hard drives, which are still relevant for secondary storage despite the popularity of SSDs. Backblaze, a cloud storage and data backup company, has published its annual report on hard drive failures for 2022.
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |