Jump to content
RESET Forums (homeservershow.com)

New BackBlaze article about Seagate ST3000DM001


ikon
 Share

Recommended Posts

The BackBlaze boys have done some more analysis about the high failure rate of the Seagate ST3000DM001 drives they've used over the past few years. You can find it here: https://www.backblaze.com/blog/3tb-hard-drive-failure/

 

 

What I got from the article is that BB thinks a good part of the high failure rate is the result of production problems brought on by the floods in Thailand. As they said though, it doesn't explain why other brands, particularly WD (who had more production tied to Thailand than anyone), didn't have failure rates this high.

Link to comment
Share on other sites

These guys need to hire someone that understands statistical analysis.

They come to conclusions even though they have no relevant data: Section on Shucking drives, you need to know where the failed drives came from!

  • Like 1
Link to comment
Share on other sites

In one table they provided, it is ironic to see that for a specific timeframe where there were only about 20% of shuckled drives compared to internal models deployed but the failure rate was in the 40%.

 

blog_seagate_ext_int.png

 

I take their research with a grain of salt, but figures this big is pretty much hard to ignore.

Link to comment
Share on other sites

For all we know in the Jan-Jun time frame all 390 externals could have failed or all 867 failure were all internals. They appear to have no idea.

They should be able to look up a serial number of a drive and know exactly what it is and were it came from. 

Looking back on their data they use a mixture of desktop and NAS drives. They mention RAID6, but they don't mention RAID controllers. If they are incorporating TLER they should use drives that support it otherwise the should not. I cant think of an valid reason they would want to mix drives.

Link to comment
Share on other sites

For all we know in the Jan-Jun time frame all 390 externals could have failed or all 867 failure were all internals. They appear to have no idea.

They should be able to look up a serial number of a drive and know exactly what it is and were it came from. 

Looking back on their data they use a mixture of desktop and NAS drives. They mention RAID6, but they don't mention RAID controllers. If they are incorporating TLER they should use drives that support it otherwise the should not. I cant think of an valid reason they would want to mix drives.

I think at one point, they mentioned used HighPoint RocketRAID controllers. you know, the Sil of enterprise controllers.

 

If they are, they're crap. Period. The controllers, that is.

To re-emphasize, BB seems more concerned about cost than quality, and I think that biases their already crappy statistics.

Link to comment
Share on other sites

These guys need to hire someone that understands statistical analysis.

They come to conclusions even though they have no relevant data: Section on Shucking drives, you need to know where the failed drives came from!

 

I confess to being a little confused by this. I thought they said all the failed drives were used in Version 2 Storage Pods. Is there something more you were looking for?

Link to comment
Share on other sites

I confess to being a little confused by this. I thought they said all the failed drives were used in Version 2 Storage Pods. Is there something more you were looking for?

They appear to be unable to identify where defective drives came from. 

blog_seagate_ext_int.png

They come to the conclusion that there is no issue with shucked drives because during Jan-Jun they had a 40% fail rate and 81% of the drives deployed were internals.

For all we know all 390 externals could be part of the 867 or all 867 could be internals. They don't seem to know.

Link to comment
Share on other sites

Basically, they haven't tracked the drives, so they have no idea WHICH failed. Just when the drives where added and how many failed.

 

It's like a vending machine. You put money in and something comes out. But you have no idea what happens inside of it.

 

But the difference here, is without that knowledge, the information they're presenting is LITERALLY USELESS.


Also, the "CSI" header for the article is dead on. It has nothing to do with fact.

Link to comment
Share on other sites

They don't use hardware RAID.  They use some sort of mdraid conglomeration.  The RAID controllers they're using are only used as HBAs, and in the case of the Highpoints they're just a bog-standard Marvell chip.  I use a RocketRAID 2680SGL as a HBA in a machine at work and it's been perfectly fine.  Again, I'm running mdraid on the drives.  

 

Though, I think the StoragePod 2.0 used the Highpoint hooked up to 5-way backplanes that have a SiL port multiplier on board, if memory serves right.  If you go back through their blog they discuss the exact hardware used in each version of the pod.

Link to comment
Share on other sites

They appear to be unable to identify where defective drives came from. 

blog_seagate_ext_int.png

They come to the conclusion that there is no issue with shucked drives because during Jan-Jun they had a 40% fail rate and 81% of the drives deployed were internals.

For all we know all 390 externals could be part of the 867 or all 867 could be internals. They don't seem to know.

 

Well, we do know that 477 (867 - 390) of the drives that were deployed from Jan to Jun were Internals; 87 more than the Externals. That means that at least 55% (477/867) of the failed drives were Internal, and that Internal drives had a minimum overall failure rate of 27% (477/1751). I don't think that's useless info.

 

I completely disagree that the BB people believe think there is no issue with shucked drives. In fact, their statement is that, regardless of whether the drives are Internal or External, the failure rate is unacceptably high. I agree with that.

 

Regardless of whether BB has tracked every last bit of data that would be scientifically ideal, I think there is valuable info that can be gleaned from the stats.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...