Hard Disk Drive Longevity with Linux vs Windows
| Author |
Message |
|
zeeone
MEPIS Rules!
Joined: Fri Jul 14, 2006 12:50 am Posts: 499 Location: Third Stone From The Sun Has thanked: 48 times Have thanks: 44 times
|
Post # 292597
 Re: Hard Disk Drive Longevity with Linux vs Windows
My HD bone pile says Windows back to MS-DOS, is the clear winner for dead ones. 
_________________ Research, Research, Research, Before You Walk The Plank. Registered Linux User # 398829
|
| Wed Jan 04, 2012 5:51 pm |
|
 |
|
JimC
MEPIS Guide
Joined: Tue Jul 25, 2006 12:47 pm Posts: 1273 Location: Savannah, GA Has thanked: 68 times Have thanks: 232 times
|
Post # 292607
 Re: Hard Disk Drive Longevity with Linux vs Windows
The WD Green Desktops drives do it, too. Basically, the drive's firmware controls it, and you can have excessive load cycle counts using Linux. From what I understand, some of these WD Green Drives are designed to park the heads after about 8 seconds of inactivity (controlled by the drive's firmware, not the OS features for power saving). But, most Linux distros will read or write to a drive about every 30 seconds. So, you end up with constant head parking and unparking with some WD Green Drives running Linux. Some of the laptop drives do it, too. I've got around 3 Million Load Cycle counts on a WD Scorpio Blue Drive in my wife's laptop now (which has run Mepis almost exclusively since we bought it), when the drive is only rated for 600,000. The Desktop drives are rated for less (the specs on the WD Green Desktop Drives show 300,000, which you can exceed in less than a year of normal use with Linux from my experience so far). See the posts I made about the counts on these drives in this thread (Desktop in one post, Laptop in the next). But, they're still chugging along, even though my wife's laptop now has about 5 times it's rated number of load cycles on it (knock on wood). ;-) viewtopic.php?f=7&t=25964&p=282565#p282565
|
| Wed Jan 04, 2012 9:34 pm |
|
 |
|
Zevon
MEPIS is cool!
Joined: Sat Dec 13, 2008 11:23 am Posts: 549 Has thanked: 80 times Have thanks: 43 times
|
Post # 292612
 Re: Hard Disk Drive Longevity with Linux vs Windows
Glad they're useful. I've run Win2000, WinXP, and currently Slackware on the machine. The HDD had power saving in automatic mode originally but I disabled it using the disk's native commands. Using 'smartctl' or possibly 'hddparm' achieves the same aim in an easy way today. The Vaio is slow by today's standards and the fan is annoying but it still does useful work. I recently saw a way to upgrade the CPU which is a bit scary looking but I might just do it anyway...
|
| Wed Jan 04, 2012 10:23 pm |
|
 |
|
m_pav
MEPIS Guide
Joined: Sun Aug 06, 2006 3:02 pm Posts: 1869 Location: New Zealand BOP Has thanked: 20 times Have thanks: 275 times
|
Post # 292618
 Re: Hard Disk Drive Longevity with Linux vs Windows
Just checked my laptop against a random windows Laptop that just happened to have a HD with 40 hours difference in operational time.
Over 3301 hours of operation time, my laptops HDD on average unloaded 54 times an hour. Over 3258 hours of operation time, windows laptop HDD on average unloaded 42 times per hour
Not worth worrying about if you ask me.
_________________ Mike P
Regd Linux User #472293 (1)Lenovo e520, i7-2640M, 8GB, 500GB Seagate Hybrid, M11-RoadblockB-64 (2) Asus M4A88TD-M, AMD Phenom II x4 955 CPU, Radeon HD 4250, 4Gb, 1.5TB, M11-64
|
| Thu Jan 05, 2012 3:34 am |
|
 |
|
JimC
MEPIS Guide
Joined: Tue Jul 25, 2006 12:47 pm Posts: 1273 Location: Savannah, GA Has thanked: 68 times Have thanks: 232 times
|
Post # 292626
 Re: Hard Disk Drive Longevity with Linux vs Windows
It looks like my Wife's laptop has averaged 120 load/unload cycles per hour over it's lifetime. I'm typing this post from it, and it looks like it's racked up another 500,000 cycles since I made my last post about it (it was showing approx. 2.9 Million in September 2011, and it's already up to over 3.4 Million, so that's another 500,000 in only 4 months, and the drive is only rated for 600,000 total over it's lifetime). But, it's still working fine (knock on wood). :-) It's been running Mepis 11 exclusively since then (I removed windows entirely when I upgraded my wife from Mepis 8.5 to Mepis 11 last year, letting the installer use the entire drive), and from the latest stats, it looks like I'm seeing a higher rate per hour now with Mepis 11. But, usage patterns probably effect it, too (so the big increase in such a short period of time may just be a coincidence). Here's it's current smartctl rersults:
|
| Thu Jan 05, 2012 9:44 am |
|
 |
|
joany
MEPIS Enthusiast
Joined: Mon Feb 12, 2007 1:45 pm Posts: 4871 Location: Penn's Woods Has thanked: 537 times Have thanks: 398 times
|
Post # 292629
 Re: Hard Disk Drive Longevity with Linux vs Windows
Thanks m_pav. That's another useful data point. I have some questions concerning my backup HDD. This is an old Seagate disk that I salvaged from a computer that ran Win98 and was later converted to Win2K before being retired back in 2006. The smartctl results for this HDD seem very odd to me: First of all, the Raw Values for some of the error rates are completely off the charts, although no problems surfaced when I ran a complete diagnostic test on the disk. I don't know what went on when the HDD was being used in the Windows machine because I wasn't using smartctl then. Taking those numbers with a grain of salt, note that the Power_Cycle_Count is 3154 whereas the Start_Stop_Count is only 7. How is this even possible? My primary Western Digital HDD show the Power_Cycle_Count = Start_Stop_Count = 2372. Those numbers make a lot more sense than 3154 vs 7. Can someone explain the strange Start_Stop_Count for the Seagate disk? (BTW, I'm not particularly worried about losing the Seagate because it's only used for backup. If it fails, I could restore a replacement disk to the same state by simply making new backups from the Western Digital HDD. I'm just curious as to why smartctl is showing these results.) T.I.A.
_________________ SimplyMEPIS is SimplyTHEBEST Mepis 8.5; 2.6.32-1-mepis-pae (Internal 2.6.32-12); 4GB RAM 2.4GHz AMD Athlon 4600+ NVidia GeForce 6150 LE; 304.51 Display Driver
|
| Thu Jan 05, 2012 10:04 am |
|
 |
|
richb
Administrator
Joined: Wed Jul 12, 2006 2:17 pm Posts: 8750 Location: Rochester NY Has thanked: 580 times Have thanks: 956 times
|
Post # 292630
 Re: Hard Disk Drive Longevity with Linux vs Windows
The start/stop count on my 6 month old laptop HD is 550. Googling this indicates to me the number is a threshold and the smaller it is the closer the drive is to failure. Any way that is my reading of the information posted. A bit counter-intuitive, but the repair guys here can verify or debunk my interpretation.
_________________ Forum Rules Guide - How to Ask for Help Link to Wiki Rich
Acer Laptop 5750G: Intel i5, 4 GIG mem, nVidia GT540M/Intel integrated graphics Kubuntu 13.04, KDE 4.10.2 M12 Alpha 2 in VirtualBox
|
| Thu Jan 05, 2012 10:14 am |
|
 |
|
JimC
MEPIS Guide
Joined: Tue Jul 25, 2006 12:47 pm Posts: 1273 Location: Savannah, GA Has thanked: 68 times Have thanks: 232 times
|
Post # 292633
 Re: Hard Disk Drive Longevity with Linux vs Windows
joany: It looks like you're looking at vendor specific attributes (like you'd get using smartctl -A /dev/sda) versus all attributes (like you'd get with smartctl -a /dev/sda). Little a is for all attributes. So, you'd probably need to find Seagate manuals to interpret what those results are telling you. Here's a page with some of the more commonly used attributes (and not all drives will support all known attributes) http://en.wikipedia.org/wiki/S.M.A.R.T. If you follow the links to some of the drive manuals, you can often see attributes supported by a given drive model. But, info on what they're for is often pretty sketchy. The issue I'm discussing is not power cycles. Load/Unload cycles are when the drive parks the heads in the landing zone after not seeing any activity for a finite period of time. Most of the discussion you see about it involves WD Green Drives on the desktop (or more often, in vendor specific forums discussing excessive load cycle counts with them using some of the popular NAS (Network Attached Storage) devices running Linux. But, a lot of laptop drives exhibit the same symptoms running Linux, where you may end up with excessive load cycle counts with them (as my wife's WD Scorpio Blue behaves running Mepis on it). Most of the WD Green Desktop drives are rated for 300,000 load cycle counts, and the laptop drives are typically rated for 600,000 load cycle counts. But, my wife's 120GB Scorpio Blue drive is still working fine, even though it's now up to over 3.4 Million Load Cycles (racking up another 500,000 in the past 4 months alone) lol
Last edited by JimC on Thu Jan 05, 2012 10:36 am, edited 1 time in total.
|
| Thu Jan 05, 2012 10:23 am |
|
 |
|
joany
MEPIS Enthusiast
Joined: Mon Feb 12, 2007 1:45 pm Posts: 4871 Location: Penn's Woods Has thanked: 537 times Have thanks: 398 times
|
Post # 292635
 Re: Hard Disk Drive Longevity with Linux vs Windows
 |  |  |  | JimC wrote: It looks like my Wife's laptop has averaged 120 load/unload cycles per hour over it's lifetime. I'm typing this post from it, and it looks like it's racked up another 500,000 cycles since I made my last post about it (it was showing approx. 2.9 Million in September 2011, and it's already up to over 3.4 Million, so that's another 500,000 in only 4 months, and the drive is only rated for 600,000 total over it's lifetime). But, it's still working fine (knock on wood). :-) It's been running Mepis 11 exclusively since then (I removed windows entirely when I upgraded my wife from Mepis 8.5 to Mepis 11 last year, letting the installer use the entire drive), and from the latest stats, it looks like I'm seeing a higher rate per hour now with Mepis 11. But, usage patterns probably effect it, too (so the big increase in such a short period of time may just be a coincidence). Here's it's current smartctl rersults: |  |  |  |  |
Your results show a Load_Cycle_Count (ID #193), whereas neither my Seagate nor my Western Digital show any values for ID #193. Should I chalk this up to the fact that HDDs used in desktops don't use this data point in S.M.A.R.T. data? Back in my corporate days when we carried around laptops, it seemed that the HDDs in those machines were failing left and right. I assumed this was due to higher temperatures and perhaps mechanical abuse. Now I'm wondering if excessive load cycles were a factor.
_________________ SimplyMEPIS is SimplyTHEBEST Mepis 8.5; 2.6.32-1-mepis-pae (Internal 2.6.32-12); 4GB RAM 2.4GHz AMD Athlon 4600+ NVidia GeForce 6150 LE; 304.51 Display Driver
|
| Thu Jan 05, 2012 10:36 am |
|
 |
|
JimC
MEPIS Guide
Joined: Tue Jul 25, 2006 12:47 pm Posts: 1273 Location: Savannah, GA Has thanked: 68 times Have thanks: 232 times
|
Post # 292636
 Re: Hard Disk Drive Longevity with Linux vs Windows
Did you try using the correct case for the options with with smartctl (it looked like you were looking at vendor specific attributes like you'd get with smartctl -A /dev/sda, versus all attributes like you'd get using smartctl -a /dev/sda). IOW, it looks like you were probably checking using a capital A (which is not going to show you all attributes).
Use a small -a to see all attributes like this (substituting the drive you want to check in place of /dev/sda)
su smartctl -a /dev/sda
|
| Thu Jan 05, 2012 10:40 am |
|
|
Who is online |
Users browsing this forum: No registered users and 2 guests |
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot post attachments in this forum
|
|