no driver

New HDD burn-in routines?

Recommended Posts

Say mr Mark of the Beast, if you label anyone who disagrees with you as "ignorant" you must have lots of friends.

Share this post


Link to post
Share on other sites

If a person's data, or uninterrupted operation, is so important that they feel that they have to run 24h of IOmeter on a consumer hard disk to test it, then they probably should not have purchased a consumer disk drive. While this does not necessarily contradict the "desktop support" level advice in this thread, my hope is that it gives a little perspective.

Disk manufacturers normally ship a free utility that can perform both quick and exhaustive tests on a disk. The only real reason to use these utilities is that, in the event that a disks fails during the test, the manufacturer will want the output of the utility before they will process an RMA. When receiving a new drive, I usually run the quick test. When receiving a used or re-certified drive, I run the exhaustive. When putting a drive into use or a mission critical application, purchase an enterprise class drive. A consumer drive should not be used for mission critical operations, even if has passed exhaustive hardware validation.

This leaves IOmeter in it's intended position, as a benchmark and performance validation tool, not a hardware validation tool. In the event that an exhaustive hardware validation tool is needed, there are several professional quality applications available, some for little to no cost.

http://www.tacktech.com/display.cfm?ttid=287

BBh

BBH

Share this post


Link to post
Share on other sites

I am sorry, but you are in for a big surprise if you think running 2 min test and commiting your data to a drive is the safe way to go.

Are you implying that enterprise drives are not failing? Or are you under the false pretense that enterprise drives are made in Venus with diamond HSAs and carbonite platters?

For most manufacturers, Enterprise drives are EXACTLY the same product as their consumer drive counterparts. The only difference is the amount of warranty and some of the QA/QC stress testing we are doing in this thread. Whatever the drive moniker is, all of drives must be subjected to same stress-testing if anyone values their data. That stress testing must at a bare minimum include reading and writing to each and every sector on the drive itself -- not random 2 min quick test. Why would anyone want to leave their data at the mercy of the drive manufacturer while he can take matters to his own hand?

From a consumer perspective, Enterprise drives are for those who has no knowledge of hardware or who has tons of money and don't know how to spend it. It is not as if we are carrying bricks on our shoulders when testing a drive. You are minding your own business and letting the software do its work at the background.

By the way, this *Enterprise* crap was invented last few years. No wonder why. Before, it had a meaning since 'Enterprise' meant SCSI, totally different manufacturing and QA methodology than their IDE desktop counterparts.

If a person's data, or uninterrupted operation, is so important that they feel that they have to run 24h of IOmeter on a consumer hard disk to test it, then they probably should not have purchased a consumer disk drive. While this does not necessarily contradict the "desktop support" level advice in this thread, my hope is that it gives a little perspective.

Disk manufacturers normally ship a free utility that can perform both quick and exhaustive tests on a disk. The only real reason to use these utilities is that, in the event that a disks fails during the test, the manufacturer will want the output of the utility before they will process an RMA. When receiving a new drive, I usually run the quick test. When receiving a used or re-certified drive, I run the exhaustive. When putting a drive into use or a mission critical application, purchase an enterprise class drive. A consumer drive should not be used for mission critical operations, even if has passed exhaustive hardware validation.

This leaves IOmeter in it's intended position, as a benchmark and performance validation tool, not a hardware validation tool. In the event that an exhaustive hardware validation tool is needed, there are several professional quality applications available, some for little to no cost.

http://www.tacktech.com/display.cfm?ttid=287

BBh

BBH

Share this post


Link to post
Share on other sites

666 did you know about this command in Vista and Windows 7?

winsat disk -seq -read -drive c

It does a disk check.

Well I've written a batch file to stress my 3 disks, I let it run for about 4 hours

:1

winsat disk -ran -read -drive l -count 2

winsat disk -seq -read -drive l -count 2

winsat disk -ran -write -drive l -count 2

winsat disk -seq -write -drive l -count 2

lol

winsat disk -ran -read -drive m -count 2

winsat disk -seq -read -drive m -count 2

winsat disk -ran -write -drive m -count 2

winsat disk -seq -write -drive m -count 2

lol

rem winsat disk -ran -read -drive k -count 2

rem winsat disk -seq -read -drive k -count 2

rem winsat disk -ran -write -drive k -count 2

rem winsat disk -seq -write -drive k -count 2

goto 1

Furthermore did you know the format command in Windows Vista or Windows 7 will actually write 0's to the drive?

I did a format with a /count 2 pass so it actually takes twice as long.

I'm now doing the final stage, re-filling all 3x1.5TB drives with RAR files and then doing a test of the RARS

if that passes, Zero fill the drives and then start loading data on them.

Share this post


Link to post
Share on other sites
666 did you know about this command in Vista and Windows 7?

winsat disk -seq -read -drive c

I don't have that in 2008 x64.

Furthermore did you know the format command in Windows Vista or Windows 7 will actually write 0's to the drive?

That is what format has been doing from DOS days. It is not Vista specific. It is the last thing i do before checking SMART values.

However, I need to see the responsiveness of LBAs to understand if I have weak sectors or if drive is about to develop surface problems. That is why I run Victoria or HDDScan. Or if the PC is offline, MHDD.

I'm now doing the final stage, re-filling all 3x1.5TB drives with RAR files and then doing a test of the RARS

if that passes, Zero fill the drives and then start loading data on them.

Why don't you just run Victoria or HDDScan and let it do all these with a single click?

Share this post


Link to post
Share on other sites
That is what format has been doing from DOS days. It is not Vista specific. It is the last thing i do before checking SMART values
No, format just reads every sector to verify it's good and clears the file allocation table. If it actually wrote 0's then recovering from a standard format wouldn't be the trivial affair it is.

Share this post


Link to post
Share on other sites

From DOS, we had unconditional format which wrote hex f6h to every sector. I am positive this is what format does now (if you are not using quick option).

That is what format has been doing from DOS days. It is not Vista specific. It is the last thing i do before checking SMART values
No, format just reads every sector to verify it's good and clears the file allocation table. If it actually wrote 0's then recovering from a standard format wouldn't be the trivial affair it is.

Share this post


Link to post
Share on other sites

I also learnt something new. You cannot recover data from any formatted (even quick) partition on an SSD with data recovery tools. Please correct me if this is wrong.

Share this post


Link to post
Share on other sites
That is what format has been doing from DOS days. It is not Vista specific. It is the last thing i do before checking SMART values.

No, format does not write zeros to the disk until Vista, I can assure you, you can unformat a 'full format' hard disk in about 30 seconds with the right tools.

However, I need to see the responsiveness of LBAs to understand if I have weak sectors or if drive is about to develop surface problems. That is why I run Victoria or HDDScan. Or if the PC is offline, MHDD.

Why don't you just run Victoria or HDDScan and let it do all these with a single click?

They don't seem to work well or at all under Win 7 64bit.

From DOS, we had unconditional format which wrote hex f6h to every sector. I am positive this is what format does now (if you are not using quick option).

Only now, it was never like this under DOS or Windows up until Vista.

Share this post


Link to post
Share on other sites
I am sorry, but you are in for a big surprise if you think running 2 min test and commiting your data to a drive is the safe way to go.

15 years as an engineer, and hands on with >100k drives. No surprises yet. I'll make real certain to let you know when it occurs. Promise.

Are you implying that enterprise drives are not failing? Or are you under the false pretense that enterprise drives are made in Venus with diamond HSAs and carbonite platters?

I am implying that enterprise drives fail less frequently than consumer drives. I (and google, and microsoft, and many other data centers around the world) also have stats to back it

For most manufacturers, Enterprise drives are EXACTLY the same product as their consumer drive counterparts.

This is false, with the possible exception of two disk manufacturers that were acquired within the past 3 years.

From a consumer perspective, Enterprise drives are for those who has no knowledge of hardware or who has tons of money and don't know how to spend it. It is not as if we are carrying bricks on our shoulders when testing a drive. You are minding your own business and letting the software do its work at the background.

This coming from a man who "would purchase an X25 simply to get rid of seek noises". Enterprise grade SATA drives have only a small premium over consumer class drives. Compare an ST.11 (consumer) and SV (pretty close to enterprise grade) series Seagate if you need a comprehensive example.

By the way, this *Enterprise* crap was invented last few years. No wonder why. Before, it had a meaning since 'Enterprise' meant SCSI, totally different manufacturing and QA methodology than their IDE desktop counterparts.

Enterprise "crap" was invented by Convex in 1985.

BBH

Share this post


Link to post
Share on other sites

Here is another shot, it's on a 2 pass format using Windows 7 and it's around the 50% mark.

2 disks have it, one less than the other, can this be fixed?! I was under the impression the Windows 7 and Vista format actually writes 0's to the disk.

Share this post


Link to post
Share on other sites

Dude i need an electron microscope to see these lines.

Run MHDD, Victoria for Windows or HDDScan and post results or see for yourself.

Share this post


Link to post
Share on other sites

I dont know. I am not familiar with what you are doing. Looks like 2 of these are faulty if from same batch. You already know tools necessary to properly diagnose. If you suspect foul-play, drop to DOS and let it run when you are not around. This way you can be sure instead of guessing.

Here check this out, about 13 hours worth of my 24 hour format - manually aligned screenshots a bit messy but it's quite clear the drop :(

http://members.iinet.net.au/~scottylans/12hour6TBformat.jpg

What do you think?

Also Victoria totally won't work on Win7 64bit I think :(

Share this post


Link to post
Share on other sites

Actually, come to think of it... WTF does that graph even represent? Are these even hard drives? If they are, they are all defective. I have 2 green craps here and their graphs look nothing like yours.

Share this post


Link to post
Share on other sites

The disks would go from 100mb a second to 50mb a second, back to 100mb a second, it's almost as if they write to platter 1, then platter 2, then platter 3, then platter 4 basically.

Whilst on platter 2 or 3 I guess, there was a rough spot and 2 disks slowed the heck down.

Share this post


Link to post
Share on other sites

This thread comes up on Google when searching on the likes of "hdd burn-in", so hence this bump and possible necro.

Would http://hddscan.com/ do as a full read/write measure (zero fill and wipe) when employing "erase mode" on maximum?

I'm meaning to use it after doing smart tests.

Reason I'm asking is because I have had a hard time finding something that would work in an Windows environment and via USB, for external USB hard drives.

Share this post


Link to post
Share on other sites

Also a real novice inquiry, while we're on it; what place does block size have in the equation, and how should the parameter be set? Does it correlate with how the disk is allocated? 1.5TB drive, unformatted from factory. Ignore this if it doesen't deserve an answer. First question is more important.

Edited by tempresident

Share this post


Link to post
Share on other sites

I'll then ask what a good solution would be for zero filling a external usb drive via Windows? Seagate tools aren't meant to work with their Freeagents and other externals, and won't even start.

Share this post


Link to post
Share on other sites
Would http://hddscan.com/ do as a full read/write measure (zero fill and wipe) when employing "erase mode" on maximum?

HDDScan seems to be a bit buggy. Nothing major but I use Victoria for Windows 4.3. Apart from the initial errors about opening the port, it works fine.

In any of these utilities, yes, erase should be sufficient.

Better option would be to use Manufacturer's utility.

Reason I'm asking is because I have had a hard time finding something that would work in an Windows environment and via USB, for external USB hard drives.

Hmm, yes we must think about this possibility also: HOW TO STRESS TEST EXTERNAL DRIVES.

I had one My Book for testing and it was a bitch to open it apart. So we must think about how to do these drives. I haven't used an external hard drive, so I cannot comment. Do manufacturers' utilities see these? I think neither SeaDiags nor WD Diags can run tests on external drives.

What about SMART data? You cannot get that off of USB unless USB chipset supports (very rare) as far as I recall.

Anyone have an idea?

Share this post


Link to post
Share on other sites
I'll then ask what a good solution would be for zero filling a external usb drive via Windows? Seagate tools aren't meant to work with their Freeagents and other externals, and won't even start.

Ah, you beat me to it. So what does everyone use on external drives?

'Block size' in what are you refering to?

6_6_6 should I be worried? My 3 new WD 2TB HDD's

No bad sectors yet.

http://members.iinet.net.au/~scottylans/2TBs.jpg

??

AbRASiON, what is the status on your drives? What did you decide to do?

Share this post


Link to post
Share on other sites

The WD's are still in service with no bad sectors - the transfer speeds are occassionally a bit bunk but overall quiet and reliable :/ Vastly superior to the Seagate filth I picked up in early 2009, lesson learnt (yes, I made sure they had the good firmware)

I'll pick up some 4TB disks and a raid cage of some sort in about 18 months I guess.

Share this post


Link to post
Share on other sites

Also cool the drive.

I read many tech forums on the internet and I get the impressions most people take care of cooling their cpu, a moderate amount their gfx card and hardly anyone talks about hdd cooling. Yet hdd's have a higher failure rate and the consequence of failure is typically more disastrous than a cpu or gfx card failing. Before I cooled my drives they were typically in the 50s and 60s in the summer and low 50s high 40s in winter. Now today they 27C and are about mid 30s in the summer.

Edited by Chrysalis

Share this post


Link to post
Share on other sites

Also cool the drive.

I am also not comfortable letting my drives get "too hot". But I also remember that one of the unexpected results from Google's drive study was that there was not a clear connection between higher operating temperatures and higher incidences of drive failure. Higher utilization of the drive also did not imply higher failure rates.

Go figure.

BTW, I believe this Google report also indicates that Google doesn't use so-called "Enterprise class" drives. Or if they do use them, they also use a large number of regular comsumer class drives. The excerpt below is from the beginning of section 2.2. I added the emphasis.

More than one hundred thousand disk drives were used for all the results presented here. The disks are a combination of serial and parallel ATA
consumer-grade
hard disk drives, ranging in speed from 5400 to 7200 rpm, and in size from 80 to 400 GB.

-irrational john

Share this post


Link to post
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