laptop harddrive Load_Cycle_Count issue

[w4tch0]

Yes, the old acpi-support script /etc/acp/power.sh doesn't get called upon thaw/resume, which causes the fact that laptop-mode does not get enabled after hibernate/sleep. And because pm-utils in hardy is quite unfinished (noone wrote proper scripts/hooks for it, noone migrated old scripts from acpi-support to pm-utils) it does not play with laptop-mode well. This could be easily fixed with addining laptop-mode start/stop to my /etc/pm/sleep.d/disk script and to the /etc/pm/power.d/disk scipt and after that we could get rid of power.sh all together.

Also note, that laptop-mode in hardy is quite an old version heavily patched (crippled ) as well (see laptop-mode website for more information).

This power management madness needs to come to an end, pm-utils needs to get polished and acpi-support needs go away

[conphara]

I didn't knew about this problem until I read about the cons of using laptop-mode. My Load Cycle Count on my SAMSUNG MP0804H (Firmware YS200-06) had a total count of 192774 . That's the result of using this laptop dual boot (XP, Opensuse-> Feisty-> Gutsy-> Hardy) for the last 2 years. Anyway I had an average count of 10 per minute. But after setting it to "hdparm -B 254" it stopped at 192774 and has not changed since thursday. Is that normal?
Now my problem is my desktop. After the 8.04.1 release in the next couple of weeks, I'm planning to relocate my hard drives so I can get Hardy instead of Gutsy in a dualboot setting, but my two SATA Seagates (ST3320620AS, firmware 3.AAK & ST3320620AS, firmware ). Using hdparm and smartctl, both of them appear to not having APM (=0). And "smartctl -A /dev/sdb" doesn't show load cycle. And trying the hdparm -B 254 command shows me an error (/dev/sdb:
setting Advanced Power Management level to 0xfe (254) HDIO_DRIVE_CMD failed: Input/output error). Using the "hdparm -I /dev/sda" command it displays "Recommended acoustic management value: 254, current value: 0".
So I can't tell what the load cycle count is, neither on Linux or in Windows (using some freeware apps). So the problem might be the firmware, which I don't like to flash to a new version (if there is one) because of warranty.

Have anyone of you guys had problems with a high increase in load cycles on an Ubuntu desktop or does this problem only occur on laptops using Linux?

What should I do?


The Seagate which is running Gutsy shows this:

ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  1 Raw_Read_Error_Rate     0x000f   111   087   006    Pre-fail  Always       -       37547317
  3 Spin_Up_Time            0x0003   096   095   000    Pre-fail  Always       -       0
  4 Start_Stop_Count        0x0032   099   099   020    Old_age   Always       -       1250
  5 Reallocated_Sector_Ct   0x0033   100   100   036    Pre-fail  Always       -       0
  7 Seek_Error_Rate         0x000f   072   060   030    Pre-fail  Always       -       20439123
  9 Power_On_Hours          0x0032   099   099   000    Old_age   Always       -       1424
 10 Spin_Retry_Count        0x0013   100   100   097    Pre-fail  Always       -       0
 12 Power_Cycle_Count       0x0032   100   100   020    Old_age   Always       -       759
187 Unknown_Attribute       0x0032   068   068   000    Old_age   Always       -       32
189 Unknown_Attribute       0x003a   100   100   000    Old_age   Always       -       0
190 Temperature_Celsius     0x0022   048   045   045    Old_age   Always   In_the_past 874774580
194 Temperature_Celsius     0x0022   052   055   000    Old_age   Always       -       52 (Lifetime Min/Max 0/12)
195 Hardware_ECC_Recovered  0x001a   079   060   000    Old_age   Always       -       223863805
197 Current_Pending_Sector  0x0012   001   001   000    Old_age   Always       -       4294967295
198 Offline_Uncorrectable   0x0010   001   001   000    Old_age   Offline      -       4294967295
199 UDMA_CRC_Error_Count    0x003e   200   200   000    Old_age   Always       -       0
200 Multi_Zone_Error_Rate   0x0000   100   253   000    Old_age   Offline      -       0
202 TA_Increase_Count       0x0032   100   253   000    Old_age   Always       -       0

[BandD]

Generally Desktop drives don't have a need to park the heads as they are not vulnerable to bumps nor do they need to conserve power. So your desktop drives probably don't have a head parking feature. So I wouldn't worry about the desktop drive.

Also regarding the laptop 200,000 cycles doesn't seem too bad to me for a 2 year old drive. But it's hard to tell where those cycles come from because you have run so many different OS's on it. The main thing to watch with changing the hdparm -B values on a laptop is the temperature of the drive. So keep an eye on that and make sure it isn't exceeding the manufacturer's specs.

Also, one other thing to note. This problem with Load Cycle Counts affects ALL OS's. It has been documented numerous times on this forum and on several other places. Yes even Windows.

[conphara]

I read about a Linux developer who said that it's better to have a hot drive than a drive which load cycle too often. My laptop drive starts with a temperature between 25-27 C and goes up to around 45-47 C (after some heavy downloading and general internet usage). Remember my laptop drive doesn't support APM (only SMART), so either way it would still be quite hot. It's only around 25-30 C IF I use the laptop-mode and the drive spins down, but I've read that that isn't a good thing long term, because a drive can only spin up and down x amount of times (like the shutdown hard drives in Windows). When I was using laptop-mode (not since -B 254) it would spin up and down constantly which made me worry.
So a good advice from me, is adjusting the setting to -B 254 can be a good thing, and in my case I could even lower it to 192 since I haven't used that many cycles.

Regarding the desktop load cycle count. Are you sure that it doesn't behave in the same way, since my drives don't support APM and therefore are set to 0 (aggressive power management). When I used the Seagate floppy tool yesterday there was a feature to test a spin down (acoustic) and it did, so the drives do support that behavior. But still I'm a bit worried since I can't read the load cycle count either in Linux (hdparm 7.5-1 with Gutsy) or Windows (freeware), I can't know for sure if the count goes up faster than I can adjust it to.

But it's rather comforting in a way that it also happens both in Windows and Mac OS. I read that the fix in mac was also the hdparm -B 254 setting. And Windows just writes to disk constantly that it never gets a chance to park. At least in Vista.

[conphara]

According to Seagate's hard drive manuals my two desktop drives don't support Advanced Power Management (APM) and Automatic Acoustic Management (AAM) features. Does this mean that as desktop users we shouldn't have to worry about this load cycle count bug?
I was under the impression that all drives counted load cycles, apm or no apm.

[frodon]

As far as i know this annoyance is laptop only. I don't think desktop computers hard drives have this feature at all.

[conphara]

Thanks,

so the morale of this story is: if your drive, a desktop or a laptop drive, supports Advanced Power Management (APM), you will experience problems with a high increase in load cycles because of APM.
And adjusting the drive via "hdparm -B 254" will solve the problem by disabling APM. The cost of this is more power and a hotter drive (just a few degrees), because it doesn't park the heads all the time. Short term solution until they come up with a proper long term fix.

In my case, I was worried that acoustic management was a big deal. First of all I didn't know what it was, http://en.wikipedia.org/wiki/Automat...tic_Management, and apparently it's only noise control vs performance. Not a huge deal.

[aldeby]

hddtemp bug seems to be fixed upstream (not yet backported to hardy though)

0.3-beta15-43
Superseded in intrepid-release on 2008-05-11

hddtemp (0.3-beta15-43) unstable; urgency=low

* Don't wake up SATA drives if not asked (closes: #479840).

-- Ubuntu Archive Auto-Sync <email address hidden> Sun, 11 May 2008 15:33:02 +0100

Thanks for the useful guide ubuntu_demon!

[I_can_see_the_light]

Hello everyone, newibe here. I have been reading about this problem back and forth but I'm not sure which solution I should use.

Smartmontools reports that my Load_cycle_count is close to 360 000 and it seems to rise by 3-6 cycles/minute. I have only been running Ubuntu Hardy for just over a month and prior to that I've been running Win Xp for 2,5 years. I have however calculated that the last 50 000 cycles have happened since I started running Ubuntu.

My question is which solution I should choose. The ugly-fix for Gutsy seemed easier to me, but I'm not sure what ubuntu_demon meant with it not being future proof. Was he talking about Hardy upgrades or when upgrading to Intrepid Ibex? 'Cause I'll probably stick with Hardy for a while since it's (basically) my first experience with GNU/Linux

Hope to hear from y'all soon

[I_can_see_the_light]

Ok, so I tried to set the APM manually by using

sudo hdparm -B 254 /dev/sda

but the output I got was

/dev/sda:
 setting Advanced Power Management level to 0xfe (254)
 HDIO_DRIVE_CMD failed: Input/output error

I googled the error message and found a command to check whether my HDD support APM: sudo hdparm -i /dev/sda | grep AdvancedPM. Apparently it doesn't, as the output I got was: AdvancedPM=no WriteCache=enabled

What do I do now? Is there any other way to force the drive not to park as often? I started synaptic and apparently laptop-mode-tools are installed. Could that be why the disk is parking so often (4.5 cycles/minute on average). I remember reading something about that somewhere, though I could be wrong.