How to fix high Load_Cycle_Count laptop drive (TOSHIBA MK6006GAH in Vaio TX1XP)?

Question

Hoping someone knows exactly what's going on here. It seems this drive has some combination of aggressive power saving settings and Ubuntu defaults that has massively increased the Load_Cycle_Count for the drive:

https://wiki.ubuntu.com/DanielHahler/Bug59695

So the drive is now so slow that it cannot boot because it takes long enough to access the data that the kernel will not recognise it properly. I'm not worried about the data on the drive, but would really like to keep the laptop functioning. There is some indication that this is possible because the figure is still low 200,000s and most drives supposedly go to 600,000. Additionally, SMART tests pass and consider the drive healthy and without errors. But the really surprising thing was when I ran mhdd...

Every single read came up red (slow) until I pressed 'R' for reset drive. I noticed the next read was normal speed, so held down 'R'. Magically the drive read perfectly for as long as I held the key BUT resumed slow (and noisy) seeking/reading after releasing. I don't think the source code to mhdd is available, so I'm not exactly sure what this means (besides, I don't know enough low-level HDD stuff either). It seems like the drive should be able to work, but is stuck trying to power save or something. There are no BIOS options on the laptop.

Does anyone know how I can stop the drive from doing extremely slow/noisy operations like this? Or is constantly resetting the drive also damanging, and only causing it to work well by luck (i.e. not a suggestion that it's fixable)?

Answer 1

A High Load_Cycle_Count is not one of the pre-fail attributes. Even though your SMART status says the hard drive is OK... chances are very good that it is not ok based on your description of noise and speed. As for running at full speed while holding down R in mhdd... I have no idea.

Hard drives that are near the end of death can make many strange noises. Constantly power cycling the drive should not damage it.

I don't trust the pass/fail guidelines that the manufacturers offer with their SMART attributes. I have seen it happen many times that hard drives are corrupting data and acting flaky even though the SMART status has a big green OK next to it. I recommend looking at the raw data counts (using HDTune in windows or smartctl in linux). From there you can make a more educated decision about the health of your drive.

Google created the largest consumer hard drive failure study... so you might want to look at their findings from 2010:

We find, for example, that after their first scan error, drives are 39 times more likely to fail within 60 days than drives with no such errors. First errors in reallocations, offline reallocations, and probational counts are also strongly correlated to higher failure probabilities. Despite those strong correlations, we find that failure prediction models based on SMART parameters alone are likely to be severely limited in their prediction accuracy, given that a large fraction of our failed drives have shown no SMART error signals whatsoever.

http://static.googleusercontent.com/media/research.google.com/en//archive/disk_failures.pdf

You also might enjoy the wikipedia article which outlines the various SMART atributes and which attributes correlate with electro-mechanical failure.

http://en.wikipedia.org/wiki/S.M.A.R.T.

The only proper way to fix a hard drive is to replace it. Fixing the one you have is something you'd only want to do if you are trying to recover data from it... and even then, you'd want to migrate everything onto a new drive. I think it's safe to say you have a broken hard drive.

Answer 2

I fixed this problem in my VAIO (TOSHIBA MQ01ABD075)

Load_Cycle_Count is not increased if you read any HDD S.M.A.R.T. option.

I use /usr/sbin/smartctl -g apm /dev/sda

#!/bin/bash

while :
  do
  sleep 3
  /usr/sbin/smartctl -g apm /dev/sda  > /dev/null
done

Answer 3

GoldMangust's answer is good; this is similar and adds some info.

This showed an ability to stop the incrementing of load cycle count (on macOS 10.12 of a Seagate Laptop SSHD Hybrid drive, Model: ST1000LM014):

while : ; do sleep 9; smartctl -A /dev/disk0 | grep "Load_Cycle_Count\|FLAG\|smartctl" ; done

Every 9 seconds seems to be often enough.

Answer 4

I found a much better solution.

• for Mac users: Install http://mckinlay.net.nz/hdapm/
• on Linux, the problem is probably long since fixed, but if not, see hdparm, its usage, and lots more about this bug by starting here
• for Windows, the linux link has a link at the end for you too.

It installs a tool called hdapm in /usr/local/bin and a LaunchDaemon plist file to run it on boot, setting all connected drives to maximum performance mode.

Given how far beyond spec my drive is, this seems appropriate for sure. Others might want to modify the plist file to set their drives to a more of a middle of the road setting.

In my case, the output is quite scary - the drive is far over spec and quite possibly near failure:

smartctl 6.5 2016-05-07 r4318 [Darwin 16.7.0 x86_64] (local build) ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED ... RAW_VALUE 193 Load_Cycle_Count 0x0032 001 001 000 Old_age Always ... 857614

.

I have smartctl installed, and by running

sudo  smartctl -g all  /dev/rdisk0,

I can see that the System Preferences setting "Put hard disks to sleep when possible" has no effect on the apm setting for this drive - it's 128 either way. I would say the drive has some unreasonable settings if a middle setting produces the result I'm seeing. Also, I wonder if even drives that come inside new Macs display/come with crazy apm behavior/settings. I would bet not, but wouldn't be that surprised if I lost such a bet.