Latest BackBlaze Drive Report Is Out
-
Sadly, like an article I wrote just spoke to, the data is too late to be truly useful. No one is buying those old drives anymore. So while interesting, it's purely historical and not really something that we can apply.
And it is not comparative. We don't know if anyone else was having issues with their drives around the same time.
-
@scottalanmiller said:
No one is buying those old drives anymore.
-
Ok, granted newer manufacture, but checkout those refurbs!
-
Same model number but presumably made years later. Does the issue carry through? They didn't test any drives since 2012 so we don't know.
-
@scottalanmiller said:
Does the issue carry through?
I plan on not finding out the hard way! I was set to buy some of those 3tb drives this weekend too.
-
The study also puts primarily failure and resilver failure together in one pool, that's not particularly useful because it is known that resilver failure rates are inflated. It's still a high number, no matter what, but my point is that the number lacks the necessary details, context and comparitives to really utilize it. It's suggestive that we should beware of buying old Seagate drives like this, but it doesn't tell us if they have been fixed or which drives made today are good or bad, only which ones were recently looking like they would be good.
-
Everyone seems to Give Seagate Barracuda Drives the most flak. I've had the most luck with them granted I use 1TB models. And they are from say 2009 at this point. I still have about 10 Hitachi 1TB Enterprise drives working fine from 2008 as well.
-
@thecreativeone91 said:
Everyone seems to Give Seagate Barracuda Drives the most flak. I've had the most luck with them granted I use 1TB models. And they are from say 2009 at this point. I still have about 10 Hitachi 1TB Enterprise drives working fine from 2008 as well.
I'm always surprised by this too. The regular ebb and flow of the fickle industry, I think. WD definitely came from behind and flipped the tables. They were the third place (in the mind) player eight years ago. They bought HGST with the best reputation and now WD is seen as the player to beat and Seagate has really been hurt. Likely it will flip back. Mostly, I think, it is marketing.
Although this one run of drives does look pretty weak.
-
WD's move to use color coded drives was brilliant. They made drive buying so much easier for the SMB segment. You can easily tell someone that they need Red, Green or Black drives and everyone understands, even if they don't understand the differences.
-
@scottalanmiller said:
WD's move to use color coded drives was brilliant. They made drive buying so much easier for the SMB segment. You can easily tell someone that they need Red, Green or Black drives and everyone understands, even if they don't understand the differences.
Yet somehow someone ends up putting Green Drive in a server where they think they need 10GB throughput for CAD, if I recall from Spiceworks.
-
@thecreativeone91 said:
Yet somehow someone ends up putting Green Drive in a server where they think they need 10GB throughput for CAD, if I recall from Spiceworks.
Well at some point nothing is going to protect you from yourself. Safeguards can only go so far.
-
Whole fleet of WD Red's at work here, they're quite nice so far. Good throughput on the NAS even in RAID5
-
@MattSpeller said:
Whole fleet of WD Red's at work here, they're quite nice so far. Good throughput on the NAS even in RAID5
We have several WD Greens here (Reds without TLER). They have been great. Two more just arrived, in fact.
-
@scottalanmiller Pretty sure TLER is responsible for the so far success we've had despite the odds
-
@MattSpeller said:
@scottalanmiller Pretty sure TLER is responsible for the so far success we've had despite the odds
It actually does very little. Only helps in very specific timing situations. I've never seen a situation with the Greens where having been Red would have mattered.
-
@scottalanmiller very interesting, thank you!
-
Not that I would use Greens instead of Reds in your case! Not saying that. Just saying that TLER rarely helps. But it does sometimes. DAC is a horrible thing to experience and TLER is about preventing DAC.
-
They often have good posts about drive reliability although they use consumer drives which are very different from the enterprise drives that most storage vendors ship.
On a side note, I worked at HP during the Thailand flooding and can tell you that the hard drive issues were real. HD prices were increasing and we could not get the quantities we needed. I saw shortages that impacted our ship first hand.
Given all these challenges it is no surprise that failure rates would go up for consumer drives. The HD companies were struggling to fill orders and they were trying to do everything they could to increase volumes. It is likely that these pressures could lead to more aggressive testing and a potential decrease in quality. Consumer drives would likely be the ones that were most directly impacted since they are typically the lowest quality.