Ubuntu May Be Killing Your Laptop's Hard Drive

wwrmn writes "There's a debate going on over at bugs.launchpad.net on whether it's the Ubuntu, BIOS, hard-drive manufacturer, or pick-any-player's fault, but Ubuntu (and perhaps any OS) may be dramatically shortening the life of your laptop's hard drive due to an aggressive power-saving feature / acpi bug / OS configuration. Regardless of where the fault lies or how it's fixed, you might want to take some actions now to try to prevent the damage."

The Ubuntu

[bazald]

FTA:

When switching to battery power, /etc/acpi/power.sh issues the command hdparm -B 1 to all block devices. This leads to extremely frequent load cycles. For example, my new thinkpad has already done well over 7000 load cycles -- in only 100 hours. That's at least one unloading per minute. Googling for "load unload cycles notebook OR laptop" shows that most laptop drives handle up to 600,000 such cycles. As these values clearly show, this issue is of high importance and should be fixed sooner rather than later.

It definitely sounds like it is "the Ubuntu" that is at fault in this case. Where is the room for doubt?

Ubuntu?

[keithjr]

If perhaps it could be "any OS" then why headline this as "Ubuntu" killing laptops? I can't find much in TFA that makes a compelling case that it isn't APCI. I'd read more but that page hurts my eyes.

Isn't this what is supposed to happen?

[zappepcs]

I mean, the OSS community at large finds a problem, and sets about to fix it... from the link:

Ralf Nieuwenhuijsen wrote on 2007-10-25: (permalink)

May i just warn ya all to NOT play the blame-game?

It does sound like it's the fault of the BIOS (and somebody should contact them).

To rescue a hard-drive in distress sounds like something that should have a high-priority (critical?).
Not because it's ubuntu's fault or the bios fault. But because Ubuntu can solve this issue _now_.

Doesn't sound like it is NOT being dealt with, it just isn't listed everywhere as critical and in the news all over the intarweb tubes.

Re:AHA! :D

[ajs]

That's why I use windows. So I don't have to wonder who the culprit is

Oh? Then you've never been caught in the "it's the graphics driver, no it's the motherboard, no it's the OS, no it's the graphics driver," loop.

This thread sucks...

[Ecuador]

I mean, if it was Windows that was destroying laptop hard drives, this would have been a legendary thread, with viciously bashing comments, insightfully (40%) funny (20%) attacks against MS, Vista drama etc.
With Ubuntu as the culprit there is some sort of "respect" that kills the potential of the thread. Come on guys, it is not Linux, it is just Ubuntu. What are the SuSE/RH/etc fans waiting for?

Not Entirely Accurate

[marcantonio]

It's important to note that this only occurs if ENABLE_LAPTOP_MODE is enabled. By default it is NOT set. From /etc/default/acpi-support:

# Switch to laptop-mode on battery power - off by default as it causes odd
# hangs on some machines
ENABLE_LAPTOP_MODE=false

Missing the point...

[msimm]

With Vista it's Microsoft's fault. With Ubuntu (or any open source project) technically, it's our fault. So if you're confused about the missing flames maybe you need to rethink what Open means.

Re:So what's new?

[Anonymous Coward]

I love *nix, but it it not ready for primetime yet

... unix isnt ready for primetime?

Re:The Ubuntu

[kebes]

Does Windows automagically disable this behavior?

Good question. The comments in the bug report speculate that Windows either completely ignores this feature, or ignores the manufacturer values and uses its own values. (In either case, what's the point in having BIOS set defaults?)

A conspiracy theorist might suggest that the hardware manufacturers are intentionally adding a feature that causes the drives to fail eventually (but statistically outside of the manufacturer's warranty period), so that people have to buy more hardware. The more likely explanation is that the manufacturer set the defaults, but didn't notice that the values were unsafe because Windows ignores them.

Getting Ubuntu to override the defaults should be viewed as a short-term solution. Ultimately the hardware manufacturers should be setting default values that will not damage the hardware. Ideally they would design safeties into the hardware, which do sanity checks and reject ridiculous values.

Re:i hate to say this but:

[LMacG]

Wasn't it HP that refused to fix somebody's laptop hinge because they were running Linux?

Ah, here it is [slashdot.org] -- sticky keys, not broken hinge, but still. You might want to give that cute gal in Canada a call back.

Re:no problem, really!

[dlZ]

I'd rather spend the extra money on Vista/Windows than have an OS that could potentially crash my hard-drive. And given the fact that laptops are all that some of us use, it's not worth the extra effort (assuming the article is right, of course).

I'd say it's more of hard drive manufacturer issue. I have 3 notebooks all running Ubuntu, and the one with a Hitachi HD had this problem, but the other two with Fujitsu HD's didn't. Luckily it took about 5 seconds to fix it. If the manufacturer set a realistic cycle this wouldn't be an issue. Ubuntu is just telling the hard drive to do it's thing, and unfortunately the hardware is set to commit suicide it seems.

Similar issue with new WD10EACS desktop drive

[mmontour]

I see a similar issue with my new WD10EACS (1 TB Western Digital "Green Power") desktop drive:

ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
    9 Power_On_Hours 0x0032 100 100 000 Old_age Always - 582
193 Load_Cycle_Count 0x0032 180 180 000 Old_age Always - 62848

I don't know the drive's rating for Load_Cycle_Count, but the scaled SMART attribute has gone down from 181 yesterday to 180 today so it does seem to be burning through its allocated cycles quite rapidly.

Interestingly, this drive does not appear to support the "hdparm -B 255" command:

mythtv:~# hdparm -B 255 /dev/sda /dev/sda:
  setting Advanced Power Management level to disabled
  HDIO_DRIVE_CMD failed: Input/output error

"hdparm -I" lists "Power Management feature set" and "Automatic Acoustic Management feature set", but not "Advanced Power Management feature set".

The system is running Debian Etch with a 2.6.23 kernel, and I'm using hdparm version 7.7. I am not using any "laptop mode" settings (at least, none that I can see).

Re:ACPI is fine

[Anonymous Coward]

ACPI is a perfectly fine standard. The problem is the motherboard manufacturers who implement shoddy bios. Almost all bioses out there fail to comply with the standard and have other bugs like this one. They don't get fixed because they just make a windows driver to work around the problem, leaving linux to bump into the bugs left in the bios.

Re:The Ubuntu

[PhilipMckrack]

I think the answer is they should both fix the problem. Drive manufacturers should issue a sane value and operating systems should check that the value is sane before using it. Same rule as accepting any data from an outside source, you tell them what they are supposed to do, but then you double check it to make sure they did. Pointing fingers at each other while customers hardware fries doesn't help anyone.

Re:no problem, really!

[EdelFactor19]

Although in some form I agree with you I'm not sure how far the trolling goes.... besides how many Laptop hard drives have hundreds of gigs of data? I've never had one that was more than 120 gig, most sold and which I encounter are 80. Yes there are larger, but I know at my workplace the corporate issue thinkpads are all on the 80gig mark; the t42p's and the t60p's.

Why is this pertinent? These are the same laptops that many college students end up with. A large amount of them have deals with ibm and now lenovo, I know RPI did and still does, I still have my 2002 t30 (third hard drive, one factory dead in 2002, one manual replaced in 2006 post warranty acts as a gentoo MythPVR sitting ontop of my cable box), and my 2004 t42 (first 80g motor died within a year, made fulltime switch to linux on it in 2006, drive got corrupted in early 07 but is back up and mostly ok although after 4 replacements of the plastic palm area from cracking the latest bunch well after warranty its days may be numbered as well). Seems odd that my ancient although somewhat upgraded (256 extra ram to get up to 512, a 120gig, 300 gig and 400 gig hard drive replacing the 40 gig drive) dell '00 XPS 1ghz desktop is outliving both of them. Granted I stopped using it for FPS and started using it almost exclusively as a linux file server (that dual boots / physical vm's into windowsXP for use of my OmniIO/Delta66 digital recording interface [i.e. one of those drives is exclusively for the resulting midi, wav and project files]) but still seems sorta odd.

back to my main point, I'd also question whether you can buy 4 laptop hd's for the price of vista and I'm by all means for sticking it to MS. Desktop hd's are heading to the ramen price segment but laptop ones are still fairly costly. More importantly there is more to the cost than just swapping the drive out and putting a new one in.. we arent talking about some random data drive that is a brainless copy over. this is the primary hard drive for a laptop and 9/10 people (if not worse) don't have a connector to connect a laptop HD to a desktop or a second laptop and thereby copy one drives contents over to another. Its bad enough to replace the batteries on laptops as frequently as we do, but ahard drive as well? not such a good sign. If these were failing desktop drives it would be more surprising but almost less important; throw raid, cheap drives and a usb enclosure at it and problem solved.

I'll concede your main point but instead of your answer, i'd simply rescind it to +3 or so :-)
lets save the flogging for whoever really is behind this

noatime & nodiratime

[redelm]

Yes, boyz'n'grlls spin-up kills hard-disks. And worse with Unix and other Linux-like OSes since they modify the directory entries each time a file is accessed (even from cache), updating the 'atime' entry. AFAIK, MS-ntfs has no such entry. Yet :)

This is a well-known performance-killer (imagine a newspool), so disks should be mount'd with the `noatime` and `nodiratime` options if at all possible. This can be done automagically by replacing 'defaults' with 'noatime,nodiratime' in /etc/fstab .

fudmuffin

[EmbeddedJanitor]

proposed new autotag for all kdawson stuff.

Re:no problem, really!

[darkmeridian]

Or you're a law school student with a laptop you keep notes on and then you write a 50 page thesis on it and then drop it or lose it. This happens a lot too, dude.

Re:The Ubuntu

[cmowire]

Almost, but not quite.

Laptop-mode solves the problem one way, without changing underlying apps, by making the write cycles bursty.

But what if you could receive a message to know that the drive had just spun up, so you could batch-commit a bunch of data and maybe do some speculative read operations?

Similarly, if you hit "Save" you really want to bypass any caching, you want it to spin up right now.

Not all disk operations are created equal.

Re:The Ubuntu

[westcoast philly]

Wow. Settle down there, son.

Okay, all he's saying is that the OS can override DEFAULT power management settings, not issue the commands to park... which is exactly what you said, albeit rather childishly paraphrased. Otherwise, how is it a DEFAULT value, rather than The Only Frickin Choice?

The ignoring bit, you seem to have misunderstood. By ignoring the manufacturer's default power saving options and overwriting them with the OS's more relaxed values, the hard drive is NOT parking itself every 5 minutes. The QA testers may not even notice the obsessive parking because windows is overwriting the default values that they themselves are trying to test.
There are APM settings you can configure in windows, too. You know in Power Options Properties? Where it says: Turn off disks after...

Does THAT make sense to you?

Oh, and what kebes posted... was a comment. Nothing less, nothing more. He was not stating this as fact by any means. In fact, he even said "The comments in the bug report speculate that Windows either completely ignores ..." He's COMMENTING on a SPECULATION... and you interpret this as him stating fact?

Allow me to quote Google:

Definitions of comment on the Web:
- remark: a statement that expresses a personal opinion or belief; "from time to time she contributed a personal comment on his account"

So "God I hate incompetent comments." is probably one of the most incompetent, and thusly ironic, comments I've heard in years. well... besides anything that's come out of GWB's gob.

Now for the disclaimer. I didn't RTFA, and I don't have any degrees or anything besides some standard CompTIA and MS certs that any monkey can get with the proper study guides and a few hundred bucks. However, I DO understand a tiny bit about how some of the world operates, and this makes sense to me. So please, make this world a better place and don't crap on someone because they offer an opinion, thought, or even a Comment!

Re:The Ubuntu

[Score Whore]

I keep seeing people say "it's the hard drive (manufacturers) fault." No it's not. You don't ask a hard drive to go into ultra low power mode if you are planning on coming back to it in just a few seconds. Ubuntu needs to pull it's head out of it's backside and stop and think about how often it hits the drive after it suggests to the drive that it's not going to be used with any frequency. This is entirely a ubuntu problem or perhaps more generally a linux problem.