Seagate Firmware Performance Differences 177
Derkjan de Haan writes "The Seagate 7200.10 disk was the first generally available desktop drive featuring perpendicular recording for increased data density. This made higher-capacity disks with excellent performance cheaper to produce. Their sequential throughput actually exceeded that of the performance king — the Western Digital Raptor, which runs at 10,000 RPM vs. the more common 7,200 RPM. But reports began to surface on the Net claiming that some 7200.10 disks had much lower performance than other, seemingly identical disks. Attention soon focused on the firmware, designated AAK, in the lower-performing disks. Units with other firmware, AAE or AAC, performed as expected. Careful benchmarks showed very mixed results. The claims found on the Net, however, have been confirmed: the AAK disk does have a much lower throughput rate than the AAE disk. While firmware can tune various aspects of performance it is highly unusual for it to affect sequential throughput. This number is pretty much a 'fact' of the disk, and should not be affected by different firmware."
bug (Score:5, Insightful)
Re: (Score:2)
Re: (Score:2)
It seems to me that the only reason people make a big deal out of this is that historically nobody is used to updating their hard drives.
Re: (Score:2)
From the article:
I'm not sure where the article got this from, but it implies that it could still be a hardware difference (e.g. slightly lower spec'd DSP in the controller?)
Not a bug, it is an OEM feature (Score:2, Informative)
Reliability (Score:5, Insightful)
Re: (Score:2)
It's the return of the old "specifications subject to change without notice". Haven't seen an abuse of that one for a while.
RAID1 (Score:4, Interesting)
On my main machine at home, I always buy my disks in groups of three drives whenever I upgrade. Two drives stay in the machine as the mirrored pair, and once a month I pull one out and stash it in a safety deposit box at my bank, and put the third drive into the machine and re-sync the mirror. That way if my house burns down / tornado smashes it or whatever bad thing that might happen, I've got a drive with my machine's image on it, no older than one month, stashed away offsite in a secure place so I can recover most all my stuff to a new machine.
Re: (Score:1)
not sure if I get you right but:
Same vendor, same product, roughly same manufacturing date? Great strategy.
Re: (Score:3, Insightful)
drive failure (Score:4, Interesting)
An experienced administrator would know there is one item in the data center everything is relying on no-one could ever think of it failing, and it will fail at the most catastrophic time you think of. It won't be all fo those 1000'thns drives failing at the same time because some plane mistook your server lights for the landing runway, It will be some cheap sprinkler, the security lock of the door, Or some manager that decides to shutdown a machine to protect it from a Denial of service attack.
If there is no such item a good BOFH will create such red button.
Re: (Score:3, Funny)
Big Red Button (Score:5, Funny)
The security for the door was malfunctioning earlier this summer, and the alarm was going off. The security guard thought the button was a shutoff switch for the security system... Luckily we had redundant servers at another location... Of course half of those didn't work...
Luckily also, this was the smaller data center at that site, so it only housed a few hundred servers... including the servers that ran many of our ATMs, and our server inventory and trouble tracking software... which didn't fail over to their backups... of course.
In addition, we had no idea where the server housing our server inventory information was... It turns out it was housed on a server called Skywalker... which we couldn't find... It turned out to be a cluster of Anakin and Amidala...
Fracking geeks.
Re: (Score:2)
Re: (Score:2)
But OTOH I also never lost all my data in the flames of my burning home, nor did I have any data damaged by a virus, nor was any sensitive data stolen from me, nor
You see, a lot of this security things are a little paranoid but I still consider the argument for not buying exactly the same stuff for backup still valid. If there is a problem with the charge changes are high that it will affect both HDs.
Since security things tend to be a little paranoid, yo
Re: (Score:2)
three out of five ain't bad (Score:2, Interesting)
supposedly it was traced to a common fault in the bearings
Re: (Score:2)
Then it was all swapped in time then, right?
Re:RAID1 (Score:5, Interesting)
It's really annoying when the following happens:
- Disk 1 dies in a RAID5 set
- Hot spare (Disk 4) comes online and starts rebuilding
- Disk 2 dies during the rebuild thrashing
- Rebuild never completes
- Put in 2 new disks
- Restore a backup
- Disk 3 fails during restoration, pulling in the hot swap (one of the new disks)
- A year later, the original hot spare (Disk 4) fails, leading to another rebuild
From my own experiences, the main culprit in these sorts of cases tend to be the bearings. Why they have a tendency to go at the same time, I have no idea. Haven't had it happen lately, but I know I'd rather avoid the problem.
Usually though, it's not the make/model/build date that is the issue, but the batch number (especially for the parts rather than the drive). Parts tend to get allocated in batches, so if you get a batch of say.... bearings, that aren't up to snuff, that batch of drives will probably fail earlier, while others (even ones manufactured on the same date) will be fine.
Re: (Score:2, Insightful)
Re: (Score:2)
Problem is, while there is a a very low average chance of a drive failure, those failures actually tend to happen in clusters. I figure there are failures in teh environmental systems allowing in contaminants, you could think of them as "Friday syndrom"; whatever it is the effect is very real. If one drive in a batch fails early, the odds another
Re: (Score:2)
Re:RAID1 (Score:5, Insightful)
And many of those are actually slower than a pure, software-only, RAID solution. Sometimes the "hardware RAID" does nothing but offload checksum calculations or other bits onto slower hardware resulting it in being a major performance hinderence rather than a performance boost. Worse yet, if your controller card dies, ALL of your data is now inaccessible. Worse yet again, there is not guarantee future hardware releases, even by the same manufacturer, will be compatible. Heck some of the really low end hardware solutions don't even provide mirrored reads, which should provide a 2x read-only performance boost.
Not all RAID is created equal. And for many, software RAID, especially for Linux users, provides a solution faster than many RAID hardware solutions, is future proof, and only costs a couple of precent in additional CPU load. Best of all, it's free and works well with LVM. In a day and age where multiple cores are common and few actually use more than one, this option doesn't have much of a downside until you're willing to look at *REAL* RAID hardware.
Re: (Score:2)
This is not always the case. With my old 3ware cards for instance, the RAID 1 format precludes using either of the drives on another controller. I have however come across some motherboard and other RAID controllers which can mirror an existing non RAID drive into a RAID 1 setup and either new drive can be used alone with a standard controller.
Heck some of the really low end hardware solutions don't even provide mirrored reads,
Re: (Score:2)
The reason for the dd to
Re: (Score:2)
Re: (Score:2)
If you look at MTBF on modern systems, any system where the higher RAID levels make sense will always have something broken. Many problems can be re-factored into a way so that the Google hardware model makes more sense than spending an exponential amount
Re: (Score:2)
Very clever! I came across an article many years ago (written in 1996) by a guy named Stephen Savage about a RAID implementation that he nicked as "AFRAID" [acm.org]. In a RAID 5 implementation, the small-write parity performance penalty can be avoided if you defer
Re: (Score:2)
Re: (Score:2)
Ha! Actually, I found it a few months after I finished my crude little batch file thing. I had used this article [lee.org] as a starting point, and just kept going with it - learned quite a bit. Probably one of the trickiest but most fun parts was setting up the keygen and transfer a minimum of user interaction, and to keep the amount of code to a minimum. It was all done with standard cygwin tools, DOS batch files, and environment variables for temporary variable storage.
Re: (Score:2)
Re: (Score:2)
Watch out for one gotcha - I've read that some (many? all?) RAID arrays are built so that they're only usable with the controller that built them. In the case of software RAID arrays, you're probably pretty safe because Windows/Linux/whatever will probably work the same way over a long period of time.
However, in the case of hardware controllers, the array format may be different between implementations. This means that you're protected against drive failure but not against controller failure/theft/burnina
Re: (Score:2)
Some of these controllers allow mirroring the contents of an existing drive and use a format where either of the RAID 1 drives can be used with a normal controller.
Re: (Score:3, Insightful)
I'd rather hot swap a failed raid drive than bring down a server to increase memory or redesign a solution from scratch in order to achieve the same performance gains. Heck, for the cost of having a coder just look at the I/O intensive code I could have bought another hard drive.
-Rick
Re: (Score:3, Insightful)
Re: (Score:2)
In which country? In some countries, high import duties and a weak local currency mean that the price of a hard drive is worth a lot more hours of labor than it would be in, for example, the United States or the United Kingdom. And across how many machines does your app run?
In the USA. Let's say I have an app that is so dependent on performance that the hit taken by running a slower hard drive is currently a show stopping issue. A new Seagate 7200.10 400GB hard drive costs right around $100 in the US, heck we'll call it $150 for OMGNeedItNow shipping or local retail price. Let's figure that there is believed to be a performance issue in the code, but that no one has worked on the project for 6 months to 1 year. Figure it takes about 2 hours for a developer to get the correct
Re: (Score:2)
Re: (Score:2)
Hardware is cheaper than labor, by orders of magnitude.
In fact, when your hardware-costs begin to escalate then most of the time it is because you were cheap (read: shortsighted)
on your labor in the past and your underqualified "write-you-many-lines-of-code-for-$50-an-hour" contractors
left you with an app that scales like a lame donkey.
Btw, swapping disks and ensuring stability is called "regular maintenance" and you should have
full-time staff
Re:Reliability (Score:5, Insightful)
Your example makes sense, but what if you've already done that? Say your app is SQL-based and does some queries that are O(n^2) complex. You've already spent $20k on a bad-ass server with RAID10, a bunch of spindles, separate transaction log drives, and as much RAM as can fit. Now, a year later, there's more records in the system and performance sucks again. Where do you go from there? These disks don't go to 11. If you want to double the performance of that $20k box, you're likely going to spend not $40k but $200k.
Once you outgrow commodity parts, if you want a 2x speedup, you'll usually have to pay 10x for it. Or wait three years. The price/performance curve is deceptively shallow towards the bottom end.
Re: (Score:2)
This is why we have things like database clusters and distributed queries. Sometimes scaling horizontally makes more sense and is cheaper than trying to scale vertically. Which probably explains why it is so popular.
Re: (Score:2)
Agreed. But if the app was never created with that in mind, it's rarely an option, and, if poorly/naively implemented, it can cause problems much worse than performance.
I think the only two multi-master database implementations I've seen so far that have been done right are Usenet and Act
Re: (Score:2)
Then the person who spec'd the system is either an idiot or incompetent. By definition you have a system which is a single point of failure which forces your system to operate below accetable minimums. With this system you should have one or more drives on hot standby or additional active disks in your array to absorb the performance hit.
Given the rec
Re: (Score:2)
-Rick
Re: (Score:2)
Linux check (Score:2)
Re: (Score:3, Informative)
Wouldn't you know that I've got an AAK disk.
Well, at least you know... (Score:3, Funny)
Re: (Score:2)
Re: (Score:2)
AAE vs AAK: It's a tie (Score:4, Insightful)
Looks like Seagate designed the new drives for servers (probably file servers) because they're really good a moving large chunks of data around, doing large reads, and large write, but not so good a a ton of little reads and writes. So, don't buy them for your desktop/workstation.
Re: (Score:1, Insightful)
Re: (Score:2)
Go read the next page
http://www.fluffles.net/articles/seagate-AAK-firm
Having a drive perform 30% slower in reading while doing a file copy, is pretty significant to me.
Re: (Score:2)
Seems to me the application benchmarks demonstrated a clear win of AAK over AAE.
Only thing the article shows AAK losing at, is working with huge files.
If your daily desktop job requires you to copy 2GB files between different harddisk, you'll probably want the AAE. If you mostly do non-benchmark stuff, AAK will probably perform better.
Re: (Score:2)
Artificial Performace limiter? (Score:1)
So does this mean they they've put some kind of speed governor on their hard drives, or am I totally misinterpreting the results?
Re: (Score:1)
Not really same drives (Score:5, Insightful)
A sad detail is that updating an AAK disk to other firmware is impossible, due to physical differences of the two disks.
(emph. mine)
Different disks have different performance. News at 11.
Re:Not really same drives (Score:5, Insightful)
Re: (Score:2)
Re: (Score:1)
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Show me a single 7100.10 that can hit those numbers for a filesize > cache size and I'll be your goatse for a week.
Re:Not really same drives (Score:4, Informative)
They were selling a USB 802.11G dongle (Model DWG-122, IIRC), one model number, *THREE* different chipsets (each requiring different drivers, only one of which had drivers for other than Windows)
Nothing on the box other than a "A" "B" or "C" in tiny print in a corner.
Re: (Score:2)
Well, that's enough, isn't it?
If these drives had such a marking, then this article wouldn't be here.
one of these cards is not like the other (Score:2)
Here's the thing. If I order a 530TX from my favorite rock-solid discount house, they will fill the order with a 530TX
iis kdawson's spell-check firmware broken? (Score:3, Funny)
Poor spell checking is pretty much a 'fact' of the browser you use when you submit articles to Slashdot, and should be affected by different editors.
Perhaps kdawson's firmware is broken?
Re: (Score:3, Funny)
The Day the Earth Stood Still (Score:5, Funny)
I have 4* 3.AAE 320GB (Score:1)
I am juggling 14 GByte pdf with a breeze (albeit acrobat reader doesn't seem to work with files larger than 4 GByte)
Re: (Score:1)
Re: (Score:1)
But yes a 200+ pages book and VERY graphic intensive one (half vector and other compressed png due to the nature of the material) will cause the size to be large.
inkscape + home brewed stuff on Linux no less.
Re: (Score:2)
Yes its a frightening thing seeing that much spinning rust with so many points of failure, but if planned for it can be damn fast for not a lot of cost or effort.
First with perpendicular recording? No. (Score:2)
Re: (Score:2)
In other words, the
BTW, 2007-2006 = 1 year, not 2 years. Even if you count month 9 minus month 4 (aug - apr) = 5 months, its still less than 1.5 years so
And which of those models were being sold before (Score:2)
Why the mixed results? (Score:2, Informative)
Sadly, we can't tell, because the author has focussed on the sensationalism of poor performance rather than asking these questions. Seems to need a few more experiments setting up
Re: (Score:3, Interesting)
http://www.seagate.com/docs/pdf/datasheet/disc/ds_ barracuda_7200_10.pdf [seagate.com]
Its only when you get down to the 80 and 120 gig sizes that the cache is reduced. And thats to save money on the production costs since the drive itself sells for less. If people want a cheaper, smaller capacity drive, they aren't likely to be willing to pay more for the 16 meg cache.
So "less
It's true (Score:5, Informative)
I have been setting up a couple of 8-drive RAID-5 arrays with these drives for some customers, and I also found out that 3.AAE drives performed much better that 3.AAK. No idea why. Seagate was unresponsive to queries about flashing the firmware and I had to replace all the 3.AAK drives by 3.AAEs.
The manufacturing country had nothing to do with it. I had some chinese 3.AAE and 3.AAK as well as taiwanese (or was that thai?) 3.AAE and 3.AAK. 3.AAE would always perform better.
The kind of testing I performed was:
Also, if you buy a retail kit (which I found cheaper than OEM at Fry's), there is no way to find out the firmware level on the box. You had to open the retail boxes to check the firmware revision on the drive itself.
One theory I have is that these drives can supposedly be configured for server or workstation workloads. It could be that AAK drives are configured for server workloads by default (unless overridden) while the AAE are configured for workstation workloads by default. I have no idea how to toggle this under Linux.
Re: (Score:2)
The most low level configuration for drives themselves I've ever heard of is to adjust the noise levels, some drives have configurable audio profiles, they can be quieter, but slower.
If you could configure firmware for differant loads, that would be really cool. But I'd imagine that hdd manufactures would be against that. If you could configure a normal desktop hdd for server con
Re: (Score:2)
Quantum gave me the files to replace the firmware image on a 5.25" Bigfoot series once. The buggy firmware was found to have a problem where if you reread the same sector with a delay between reads corrupted data would be returned.
Re: (Score:2)
I have two questions that are going to sound a bit trollish, so I apologize in advance, but you sound like the right person to ask. I really wouldn't give a damn about this except that I buy a hard drive now and then and don't like to buy from companies that jerk people around.
According to the article, the AAK drives perform a bit better than
Easy solution (Score:1)
Re: (Score:2)
Whereas, I'm sure they are happy to cannibalize Western Digital's sales of 10k RPM SATA drives.
Speculating is fun (Score:1)
Very Similar problem with the Samsung HM250JI (Score:2)
Turned out Samsung had a couple of different firmware versions on shipping drives, and it is possible to burn new firmware to a CD and boot from it under OS X to
Seagates, and Linux HD optimization (Score:4, Insightful)
Also, if you're concerned about Linux block device performance, look at the various kernel tunables. On a single drive, such as those Seagates, I can get extra ~10MB/s. On RAIDs and LVM volumes, the differences can be much higher-- more than twice as fast, in some cases. There are a few parameters that make a difference, and many values you might want to try for each. I have a script iterate through the various permutations, running IOZone on each, so I can see what does best for read vs. write and large vs. small file performance. But I can't release it just yet (employer makes 100% of income from Open Source; employer hates Open Source). Anyway, somebody out there can do better than I, I'm sure :)
This discusses the tunables you'd want to check: http://www.3ware.com/KB/article.aspx?id=11050 [3ware.com]
Note that these do NOT apply only to 3Ware controllers. And the differences in performance can be massive.
Re: (Score:2)
So of course I went spelunking who your employer might be... No luck, but I got a new sig.
Re: (Score:2)
Don't like it. Makes me think of paper cuts.
Re: (Score:2)
Note that these do NOT apply only to 3Ware controllers. And the differences in performance can be massive.
Funny; 3ware controllers are slow as shit. Buy an Areca, and enjoy performance that is TENFOLD that.
I've seen twelve drive arrays on 3ware cards struggle to do better than 25MB/sec. Areca cards easily hit well over 200MB/sec from just a handful of drives (ie, take single drive speed, multiply by # of drives, subtract a little.)
Re: (Score:2)
Perpendicular Recording Overview Video (Score:4, Funny)
Observation from the "inside" (Score:5, Informative)
A hard drive is a very complex subsystem inside your computer, more complex than many people realize. A hard drive contains one or more CPUs, memory, firmware, and dedicated hardware devoted to the functions of storing and retrieving data.
There is no single "right" way to draw the line between what is firmware and what is hardware in a hard drive. Algorithms could be coded in VHDL or Verilog and synthesized into the silicon, or they could be compiled in C (or hand coded in assembly) and be embedded in firmware. Each drive company has their own philosophy for where to draw the line.
Some drive companies choose to implement only fundamental functions in silicon, and implement everything else in firmware. For these companies, comparing their firmware to the BIOS in a PC is a poor analogy. A better analogy would be to compare the firmware to the operating system.
In a system with "lite" firmware, the firmware typically would be responsible for configuring a few control registers and buffers, and then the hardware would take over. But for a system with "heavy" firmware, the firmware behaves much more like a kernel. Data is not going to be moved in or out of buffers, or be sent to and from platters, without the active involvement of the firmware scheduling and ordering that activity.
The author of the OP wrote "it is highly unusual for (firmware) to affect sequential throughput". The author is wrong. In a system with "heavy" firmware, all performance is highly dependent on the firmware. It can easily make the same difference in performance as you would see running Windows 95 v. Windows XP v. Windows Vista v. RH 7.2 v. RHEL 3.0 on the same PC hardware.
I do not know if the Seagate drive in question is a "heavy" or "lite" firmware drive, but I do know that the assumption that firmware takes a minor role in hard drive performance is mistaken.
Seagate's 7200.10's have been hit and miss (Score:2)
This would concur with the news post here, I'd say it's the AAM (or is it AMM?) accoustic management stuff.
On top of this there's been more reports of faults than there was with the 7200.9's. (although nothing deathstar esque) If you go on the newegg.com feedback section for the varying 7200.10's you'll see a surprising variety of votes, yet newegg is traditionally filled with 'fanboy' response
just got off the phone with Seagate (Score:2)
Re:why do girls always abandon me? (Score:4, Funny)
Re: (Score:2)
Re: (Score:2)
410AS is a single platter, vs 2 platters. It's also a much thinner drive.
I just recently purchased 2 of the 410AS drives to add to my RAID5 set (which has mainly older 250gb 7200.8's). Firmware on these is AAA. Performance seems to be great, though I didn't benchmark them separately.
Re: (Score:2)
Meaning of N (Score:2)
In the