laptop hdd issues with linux ?

M4rc0 · #1 17th July 2008, 08:43 PM

I just got my dell xps 1530, a week ago, and i've been noticing these "clicks" that it makes and wich increases it's Load_Cycle_Count on hdparm each time.

I browsed about it and got some information that i'd like to share with some people:

(quotes from Archlinux forums)

Quote:

It appears to be the official policy of Ubuntu (citation/disagreement, anyone?) that by default, Ubuntu should not adjust any power management settings of the harddisk. Unfortunately, this policy has two negative effects: It leaves quite a few people with broken hard drives that would otherwise not be broken, and it quite simply makes people who love Ubuntu feel neglected. This issue has been going on a long time.

The problem appears to be that some manufacturers' defaults are too aggressive and that Ubuntu might cause too many unbuffered disk accesses -- the combination of which can cause over a thousand parks a day on some systems.

(...)

Quote:

From further reading of linked material, the claims and concerns are basically these:

- HDD manufacturers are setting aggressive defaults. Some are saying the balance is struck in favor of longer battery uptime at the expense of HDD wear. The theory goes that the former helps the laptop companies promote the battery uptime as a selling point, and the latter helps HDD manufacturers sell more disks;

- Other OSes (Windows, for example), override the defaults resulting in less aggressive settings, and theoretically, longer disk life when Load_Cycle_Count is compared to stated life expectancy specs for the HDDs.

Ubuntu is know for that problem, but Fedora has it too.

I confirmed that not only the Load_Cycle_Count increases every "click" sound:

Code:

[root@localhost ~]# smartctl -a /dev/sda | grep Load_Cycle_Count
193 Load_Cycle_Count        0x0022   097   097   000    Old_age   Always       -       6950

It was less than 6900 before, and it keeps increasing!

And my temperature seems to be high too:

Code:

[root@localhost ~]# smartctl -a /dev/sda | grep Temp
190 Airflow_Temperature_Cel 0x0022   057   045   045    Old_age   Always   In_the_past 43 (Lifetime Min/Max 37/43)
194 Temperature_Celsius     0x001a   043   055   000    Old_age   Always       -       43 (0 24 0 0)

This is currently an issue with many laptop models.

Here how to fix it,also how to detect if you suffer from this,

This is very important for those that use laptops and i tough i could share as i didn't see anything like it here in fedoraforums.

Also, i would like to know here if Fedora could have a different aproach than that "ugly fix" mentioned as we don't seem to have the paths:
/etc/acpi/resume.d/
/etc/acpi/start.d/
/etc/acpi/ac.d/
/etc/acpi/battery.d/

Load_Cycle_Count 7017 in the end of writing this post.

Check your Load_Cycle_Count with:

Code:

smartctl -a /dev/sda | grep Load_Cycle_Count

Check your hd temperature with:

Code:

smartctl -a /dev/sda | grep Temp

leigh123linux · #2 17th July 2008, 09:12 PM

Please don't mark you post as important!!

4.

Quote:

When Asking Questions

1. Read The Unofficial Fedora F.A.Q. (Frequently asked questions) and Search FedoraForum.org before asking questions. Search the Fedora mailing lists and give google a whirl as well. There's a good chance your question has been answered before.
2. Details, details, details! Don't be brief. Include your Core release number, the architecture of your computer (i.e. AMD 64 or a Pentium 4, or in the case of video cards NVIDIA model or ATi Model), the version of the component you're having trouble with and everything you think might help. If you're using 3rd party RPMs explain where you got them from. Attach screen shots or other files if you think it might be useful.
3. Don't talk yourself down. Don't call yourself an idiot, we have all been newbies at some point. Don't add things like this to your posts:
Quote:
I know I'm an idiot but I just can't figure this out..
No question is too stupid.
4. Don't mark your question as important or critical. All questions asked are equally important.
5. Come back to the forum even after your question has been answered and give your thanks to the people who helped you. Try to generate that good, warm, fuzzy feeling amongst your peers.

http://www.fedoraforum.org/?view=guide

M4rc0 · #3 17th July 2008, 09:20 PM

sorry!

(well, but this is important...)

Omni-Tom · #4 18th July 2008, 03:59 AM

Hi I have a dell xps 1530 I bought in January with a 160Gig hard drive which was way to small for my liking of a dual boot Linux/Vista. Any ways ended up buying a western digital 320Gig 5400rpm hard drive as it was one sale for dird cheap as far as laptop hard drives go and do not hear constant clicking though it does click once and a while. Also my vista score for the hard drive went way up. I'm mentioning this as it might not be a problem specific to xps 1530's but may be 1530's with a specific hard drive make.

This can be worse as to fixing this problem as it may be specific to a computer type and a specific hard drive make

M4rc0 · #5 18th July 2008, 09:29 AM

Maybe.

But i am surely worried about this, yesterday the temp got up to 47 C (!)

I turned off the computer, altho i have the one year support by dell and this is a week old notebook, i don't want to risk it.

With me the ugly fix with 254 worked better as 255 made my hdd heat up.

Now i'm just wondering how to make that permanent under Fedora. Does anybody know how?

M4rc0 · #6 18th July 2008, 12:46 PM

Damn my hd just got 49C now!

What should i do ???

Nobody else feels this problem? It's a major issue

M4rc0 · #7 18th July 2008, 05:00 PM

Ok maybe i'm the only one with this issue? Anyway, here's a good explanation of the bug:

Quote:

Symptoms of this bug are:
* Frequent HD clicks -- more than one per 3 minutes while idle, louder than the typical access sounds. Often more than twice per minute. On some disks, the click is very quiet
* Rapidly Increasing Load_Cycle_Count as displayed in the final number in "sudo smartctl -a /dev/hda | grep Load_Cycle_Count" (where /dev/hda is replaced with your own hard disk device)
* Early hard disk failure never stay parked, due to very frequent disk activity. Thus this cycle occurs often, thus wearing out the drive, and any comparative benefit is negligible (whereas, if the-- some disks are cut down to less than a year of actual uptime.

The problem is only present due to the existence of *all four* of the following factors:
* Hardware is set (default or otherwise) to aggressive power management, causing heads to park. (default behaviour of many drives and often the only user available type of power management)
* Disk is touched often, causing heads to unpark. (default behaviour of many distributions)
* Drives are spec'd to a limited number of these cycles. (600,000 is the most common, although some may be spec'd higher or lower).
* The OS not setting disk APM variables according to current disk access pattern.

Reasonable Limits / Criteria for a fix:
* There should be fewer than ~15 load cycles per hour, except during heavy usage while on battery.
* This provides a life expectancy of over four years, which is reasonable for a hard disk.

And i'm trying the fix in this link, wich seems to work *feeling calmer*

Omni-Tom · #8 18th July 2008, 06:40 PM

Not shur on this but think the 1530 has a shared fan in that the fan cools the cpu and also the rest of the computer. The problem with this is that if the fan is controlled by the cpu temp that it may ignore the other parts of the computer.

This has been an issue with the graphics cards with them running hot nad possibly failing and may now be a problem with the hard drive or hard drive fix as in running more than then the xps was designed and tested with.

I have tested temps in vista and Fedora as I was trying to see if there was a cooling fan problem but the fan is so quit I was in a library when I finally heard it running lol.
Generally I found that the graphics card runs any ware from 5 to 10* hotter than the main cpu temp. My hard drive right now is at about 43* which seems to be a bit above that in Vista by 3 or 4*.

I also found with the 2.2ghz processor and 265megs vid that things can heat up really quickly and change very fast. This may be another problem in that. So for example say the under a heavy load the processor with heat up so fast that when the fan spools up the temp is way up before cooling starts. this can also cause problems with the graphics card and hard drive in that under a heavy load the temps on both may peak very high and cool down slower than the cpu. I also found that the cpu also dumps temp really well which again may cause a problem as the cooling system may spool down before affectively cooling the gpu and hard drive.

I just started using Fedora -64 after using Ubuntu for 4 years and this is the first linux distro on this laptop but from what I have read Ubuntu has worse heat problems with fedora and 64-bit is not recommended for Ubuntu yet-- as for fedora running 64 bit no problems as of yet.

Also I have the big bulky 9 cell battery which hangs about an inch under my laptop, this has a weird affect in that it props it up on a slight angle which also seems to help cooling with airflow under the laptop. You might be able to prop it up a bit and should see a huge difference in cooling.

I just started using Fedora as of yet and am still trying to get the fan speed thing working as I am not showing any fan sensors as of yet but am planning to see if I can make the fan a bit more aggressive pertaining to cpu temp an hopefully get the fan to run longer to better cool the other hardware.

M4rc0 · #9 18th July 2008, 07:00 PM

Omni-Tom, you might have a good point there, But i have to disagree the possibility of beeing an exclusive XPS 1530 problem.

In this link you can see the list of the affected laptops.
I know DELL has a major part in that list lol but you can see Thinkpad and HP as well. Maybe the new slim models are having a bad designed fan? i don't know.

What i do know, is after the fix i did, my hd cycles stopped and my temp won't go further than 43C wich makes me happy because i was getting 47, 49 and up.. The cycles were crazy, it was clicking frenetically.

I'm impressed i didn't see any topic or theory about this in the fedora community, as well as with this topic you can tell that apparently nobody cares that your hard disk might be dying within an year of use due to this linux bug.

But well, if anybody get's here, it's documented. We do have a solution for Fedora as well

Omni-Tom · #10 18th July 2008, 07:18 PM

I hear your pain I remember reading about this on the Ubuntu Forum around January so this problem has been there for a long time as in linux time. Also I agree this is more serious than what is being done about it to an extent.

Remember reading on the Ubuntu Forum in that the hard drive life was stated as cycles in that each time it clicked that the life of the hard drive was compromised, not sure if this is totally true but according to that statement someone running this problem for months even if the hard drive does not fail it will shorted the life cycle of the hard drive overall.

The part I agree with is that a user should not have to run his computer with this problem with months and find out about it from a search or reading it later on a forum. Even if not patched there should be a warning be it on first boot or where a user will see it.

Omni-Tom · #11 18th July 2008, 07:36 PM

Hi did the checks above but have no idea what its stating. I just put this hard drive in about two weeks ago.

Code:

[root@localhost ~]# smartctl -a /dev/sda | grep Load_Cycle_Count 193 Load_Cycle_Count        0x0032   199   199   000    Old_age   Always       -       5778

Code:

[root@localhost ~]# smartctl -a /dev/sda | grep Temp 194 Temperature_Celsius     0x0022   104   098   000    Old_age   Always       -       43

The count is up like 1k from a few days ago.?

M4rc0 · #12 18th July 2008, 07:45 PM

Quote:

The part I agree with is that a user should not have to run his computer with this problem with months and find out about it from a search or reading it later on a forum. Even if not patched there should be a warning be it on first boot or where a user will see it.

Totally!!

and your hd looks good in my opinion, just pay attention when it makes the *click* and then you run the smartctl to see if the Cycle increased by one. And also the temperature. Is your hd beeing noisy? i can barely hear mine now

(this laptop surely is quiet, confirming now after i read some reviews before buying it, and now that my case is hopefully solved)

Keep doing that for some minutes, if your temp stays on 43 you're good, and your Cycles are pretty low, just so you know i'm having here 7360 on Cycles, but that's because of those frenetic clicks that stopped now. Some people said it should be around 6000 ?

The temperature starts with 43 and can increase slowly, wich is normal. If your notebook is on a table, it can go to around 45 and stay.

What isn't normal is reaching 49 like mine did before and without the fan on.

Omni-Tom · #13 18th July 2008, 08:30 PM

K tried something a bit different after reading a post on one of the bad things can happen links from Disk_Power_Management I am trying to do some timed tests watching the clock and recording the cycle counts per one min.

Code:

Load_Cycle_Count :test=1 for 60seconds =3
Load_Cycle_Count :test=2 for 60seconds =4

Omni-Tom · #14 18th July 2008, 08:53 PM

Doing some reading and seeing some interesting stuff. This problem is bad but also a sheep in wolves clothing,,, just started reading and there are huge amounts of tricks and information available that may be useful for laptop users in general.

For example found this snippet on lesswatts.org/tips/disks

Quote:

Filesystem atime

The POSIX standard requires operating systems to: keep track of the last time each file was accessed by an application or the user, and to store this timestamp as part of the filesystem data. This timestamp, called atime. atime, can be useful in finding out which files are never used (to clean up the /tmp directory for example) or to find out if a file has been looked at after it changed (this is sometimes used by email applications to warn about newly received email). A significant downside to atime, is that every time a file is accessed, the kernel has to write a new timestamp to the disk, at least after a few seconds of activity. These disk writes will keep the disk and the link to the disk busy, and that costs both performance (the disk is prevented from doing something else) and power (with ALPM, the link to the disk needs to be powered back on, and if the disk was in low power mode, it would need to go to full power as well). You can disable atime entirely using this command:

mount -o remount,noatime

Filesystem relatime

Because few programs use atime, disabling it always and for everyone is not very practical. The Linux kernel has a feature called relatime, which is an effective compromise between having some of the information that atime provides, without having the disk time updated every single time.

relatime works by updating the atime field on disk only if the file hasn't been accessed since the last time it was accessed, (to provide the new email detection capability) or when the last access was more than 1 day ago (to help programs and users clean up unused files in the /tmp directory).

A discussion of the relatime option can be found at http://lwn.net/Articles/244829/.

Older version of the mount command do not support relatime, but you can try the following command to see if your mount program and kernel support relatime:

mount -o remount,relatime /

Another option is to use the nodiratime option instead of noatime or relatime. nodiratime will not update the atime field on directory inodes. This allows you to continue to update file access on read, for programs which require it, but prevents updates for just reading a directory, such as with ls.

mount -o remount,nodiratime /
/

A discussion of the relatime option can be found here

Quote:

The VM writeback time The VM subsystem, in the Linux kernel, buffers writes to files that applications perform for a period of time. This caching allows the kernel to group consecutive writes into one big write, and to generally optimize the disk IO to be the most efficient. The kernel, by default, will start writing out data to disk after 5 seconds, so, if a power failure or kernel crash happens, at most, 5 seconds of data would be lost. If your machine has a reliable power supply (for example because you have a UPS or because your laptop has a battery) and doesn't crash much, you can increase this time so that the VM subsystem can do a better job optimizing the IO, as well as having the disk (and the SATA link when using ALPM) in lower power states for a longer period of time. You can check the current value of this writeback time with this command: # cat /proc/sys/vm/dirty_writeback_centisecs 500 This example shows that the timeout on this system was 500 centiseconds (or 5 seconds in human terms). To increase the timeout to 15 seconds, you can issue this command: echo 1500 > /proc/sys/vm/dirty_writeback_centisecs

I found this interesting pertaining to laptop mode.

Quote:

For several years, the Linux kernel has had a "laptop mode" for the IO subsystem. When laptop mode is enabled, the kernel will try to be smart about when to do IO, to give the disk and the SATA links as much time as possible in a low power state.

Code:

You can see if laptop mode is active with this command:  
# cat /proc/sys/vm/laptop_mode 5 

 If the content of that file is 0, laptop mode is disabled. In the example here, laptop mode is enabled.  To enable laptop mode, you can use this command:

  echo 5 > /proc/sys/vm/laptop_mode  Several distributions also include the laptopmode scripts as part of their standard installation. These scripts enable laptop mode automatically, in some cases.

Did the laptop mode test and came back as zero but not even shure if this is even used anymore?

Most of this is from http://www.lesswatts.org/tips/disks.php which is on laptop power saving and there is a lot of info pertaining to hard drive access and limiting writes in general which may help overall problems but not sure. Less writes may lower cycles a bit as in restarting after parking.

M4rc0 · #15 18th July 2008, 10:30 PM

Nice info Omni!

We should get started with an m1530 group and do all this"scripts" and "fixes" in a single sh file. That would be neat.

I also have a tip to share:

Improve system performance by moving your log files to RAM

Quote:

The Ramlog project lets you keep your system logs in RAM while your machine is running and copies them to disk when you shut down. If you are running a laptop or mobile device with syslog enabled, Ramlog might help you increase your battery life or the life of the flash drive on your mobile device.

I think that is a must.
And i will enable laptop mode as well.

I will read your post carefully again later on.