Hard drive Failure Rates
- Thread starter Bozo
- Start date
sechs
Storage? I am Storage!
Is WD being hands-off with the HGST drives? Why are they continuing to not inherit WD's poor quality?
Mercutio
Fatwah on Western Digital
Buck said it's operating independently. I trust his insight on the matter but I don't know why they wouldn't be replacing everything they do with HGST tech.
Stereodude
Not really a
The data doesn't paint a very good picture of Seagate, yet they keep buying Seagate drives. Lots of people use their reports as the justification for not buying Seagate.
Mercutio
Fatwah on Western Digital
The most current report makes WD look worse. Once the 3TB "DM" Barracudas were removed from their pods, the rest of Seagate's drives fell in at approximately average levels of reliability, including the other 3TB Seagate models. If you read the whole post, they basically say that the Seagate drives still make the most sense because that 15% lower unit cost matters more for their needs than 50% greater per unit reliability (compared to HGST) based on the sheer number of drives they're using.
Howell
Storage? I am Storage!
Some manufacturers may not be able to meet their volume demands either.
Mercutio
Fatwah on Western Digital
Yeah. They actually say that it has a lot to do with OEM availability from their channel partners.
Stereodude
Not really a
They're not buying that many drives. They said they bought 16k drives last year? I can only imaging that Dell, HP, Levovo, etc buy orders of magnitude more drives. I can't imagine that the various HDD makers don't have the capacity to support their volumes. I'm sure they're buying from different channels though.
Mercutio
Fatwah on Western Digital
Consider that after the floods in 2011, these guys were going around to Sam's Club and Costco to buy up local stock so they could de-shell cheap external drives. I'm guessing they're not big enough to get the breaks they want/need directly from drive vendors or top-tier suppliers but too big for retail or consumer supply. I also wonder if publishing their reliability info strains any relationship they might have with drive manufacturers.
/conspiracy theory on
Or perhaps a "strained" relationship led to the publishing of their data? Or they want to get a better deal for HGST drives so massage the stats?
/conspiracy theory off
Stereodude
Not really a
That's when they were small. I'd like to see a graph of their drive buying numbers per year. I'm sure they could get drives direct now.
mubs
Storage? I am Storage!
I do not see HGST 6TB NAS drives listed in their data. I have several in my workstations and they've been fine but I'm a tiny sample.
I have one and it is rather oddly noisy. I'm concerned that an octo-RAD would be terrible.
sechs
Storage? I am Storage!
Quarterly update:
https://www.backblaze.com/blog/hard-drive-failure-rates-q2-2016/
https://www.backblaze.com/blog/hard-drive-failure-rates-q2-2016/
I never understand their logic. Does the Backblaze just buy whatever is on sale that week? Most of the cheap drives are not designed for ser multi-disk environment with vibration detection and compensation, so how do they account for that?
sechs
Storage? I am Storage!
They cost half as much enterprise drives and fail at a less than 25% (total) rate.
They've built their system around redundancy and can take the drive loses. Considering the number of drives that they buy, it's an acceptable trade-off.
Chewy509
Wotty wot wot.
Necro thread...
Backblaze reliability for Q1 - 2019.
https://www.backblaze.com/blog/backblaze-hard-drive-stats-q1-2019/
Backblaze reliability for Q1 - 2019.
https://www.backblaze.com/blog/backblaze-hard-drive-stats-q1-2019/
time
Storage? I am Storage!
I wondered how 99 drives recorded 12662 days of operation in a three month (90 day) period. You have to add back the figures from the Removed table to the closing Counts.
It's interesting how many 12 and 14TB drives seem to suffer from infant mortality, based on the Average Days to failure. The pattern suggests that after that, they're incredibly reliable.
It's interesting how many 12 and 14TB drives seem to suffer from infant mortality, based on the Average Days to failure. The pattern suggests that after that, they're incredibly reliable.
Chewy509
Wotty wot wot.
Another necro thread rise...
Backblaze reliability for Q1 - 2023.
arstechnica.com
Looks like anything 10TB+ in size, expect it to die in 3yrs...
Backblaze reliability for Q1 - 2023.
HDDs typically failed in under 3 years in Backblaze study of 17,155 failed drives
Seagate still stands out.
arstechnica.com
Looks like anything 10TB+ in size, expect it to die in 3yrs...
sdbardwick
Storage is cool
I don't think that is correct takeaway. Plus the ArsTechnica headline is grossly misleading. They should have said something like "Of drives that failed, they typically failed in under 3 years." to be less misleading.
Also, "The drive with the lowest AFR[ Annual Failure Rate] (0.28 percent) and at least 2.2 million drive days is Western Digital Corporation (WDC)'s 16TB WUH721816ALE6L4 (Backblaze has 14,098 units)."
No, it's just showing that the most common age of failures is less than three yaers. That does not count the vast majority of drives that are still working. Their aggregate AFR is 1.40%. Other than the one 14TB Segate drive with higher than typical AFR, there is not a lot to practically separate them.Another necro thread rise...
Backblaze reliability for Q1 - 2023.
HDDs typically failed in under 3 years in Backblaze study of 17,155 failed drives
Seagate still stands out.
Looks like anything 10TB+ in size, expect it to die in 3yrs...
Exactly, and the 16TB drives probably have limited hours as they would be newer drives.
Mercutio
Fatwah on Western Digital
From what I gathered, the reality is that Backblaze loses something like 1.5% of the drives it buys within three years. That's an astonishing level of overall reliability. Seagate drives are a bit less reliable, but they also cost less, so the relative value from their side very close either way.
Mercutio
Fatwah on Western Digital
Another year, another Backblaze data dump. 10 and 12GB drives seem to be problem children. 16TB seems to be a reliability sweet spot.
It's hard to extrapolate newer, larger drives to compare with older, smaller drives. I'm surprised that they are using so many 10TB drives with an average of 6 years of service and some drives are even older. Is that normal in the storage industry like Amazon and the Google or just because Backblaze is a smaller company? At some point the economics of the supporting hardware and facility infrastructure would make small drives undesirable.
FWIW My 8x10TB Seagates (NM0016) from as far back back as 2016 are all fine, but they don't rotate all the time since the NAS spins them down after 60 minutes if it is even on.
FWIW My 8x10TB Seagates (NM0016) from as far back back as 2016 are all fine, but they don't rotate all the time since the NAS spins them down after 60 minutes if it is even on.
Last edited: