Hard drive issues

[FrankBarmentlo]

I am fairly new to linux, and ran in a few problems since installation on my laptop
most of them are just solved by doing some research, but one of them really bothers me: my harddrive reaches temperatures between 45~50 degrees celsius, on idle.
it also feels like the disk is spinning at full speed, which shouldn't be the case.
technical details:
hard drive is a st750LX003,
laptop is a samsung rc730-s02nl(if relevant), dual-boot Kubuntu and win8

this is the top5 results in Iotop:
338 be/3 root 0.00 B 260.00 K 0.00 % 0.58 % [jbd2/sdb4-8]
6699 be/4 frank 0.00 B 24.00 K 0.00 % 0.02 % chromium-browser
8742 be/4 frank 64.00 K 52.00 K 0.00 % 0.02 % konsole
6689 be/4 frank 0.00 B 92.00 K 0.00 % 0.01 % chromium-browser
6691 be/4 frank 0.00 B 44.00 K 0.00 % 0.00 % chromium-browser
ah, maybe this is a nice addition:

this is the disk activity from 30 minutes.. disk temp raised from 28 to 55 degrees celsius within that timespan..

my teacher, who knows Linux/Ubuntu better then I do, thinks this is a driver issue.. but how can I solve it, if there are no linux drivers available?
edit after a lot of thinking and researching:
could it be an issue with Ext4? one of my teachers(ubuntu-user), told me it could be the way ext4 works, and going back to ext3 could solve my issue..
it doesn't happen in windows 8(NTFS), at all.. no changes, no high temperatures.. just 30 degrees celsius..

I am not sure what to post more, so let me know if more information is needed.
kind regards,
Frank

[oshunluvr]

While I agree that 50c is too hot for normal idle temp, it's not out of range. I'd be a bit concerned too though. My older drives run at 39c.

No need to go back to EXT3. If you're worried about activity, you can turn off time stamping and or make other changes to reduce access. For that matter, you can use any other disc format including ntfs but it won't change the amount of access much. It's more likely something else is causing the extra access like the IO scheduler. With the hybrid drive you have, I don't know which scheduler is best. CFQ is default, but Deadline or NoOp is usually recommended for SSDs.

BTW, there are no "drivers" for hard drives that I've ever heard of. The kernel takes care of that for you. No offense to your teacher, but he/she should know that if they know much about Linux.

I wonder if your drive is properly reporting it's temp. Does is start out at 30 and then climb to 50 or start out at 50? Open a terminal and type

sudo hddtemp -D /dev/sda

and post the results.

[FrankBarmentlo]

it start at 28, climbs to 50~55 within 30 minutes, on idle(just the login screen)

================= hddtemp 0.3-beta15 ==================
Model: ST750LX003-1AC154

field(1) = 0
field(3) = 0
field(4) = 171
field(5) = 0
field(7) = 117
field(9) = 235
field(10) = 0
field(12) = 237
field(184) = 0
field(187) = 0
field(188) = 0
field(189) = 0
field(190) = 48
field(191) = 0
field(192) = 72
field(193) = 22
field(194) = 48
field(195) = 0
field(197) = 0
field(198) = 0
field(199) = 0
field(254) = 0

[oshunluvr]

Also, lets look at it's smart status. In a terminal, type this:

sudo /usr/sbin/update-smart-drivedb

then type:

sudo smartctl -i /dev/sda

and post the results.

[oshunluvr]

I found this too:

http://forums.seagate.com/t5/Momentu...hot/m-p/161426

[FrankBarmentlo]

Also, lets look at it's smart status. In a terminal, type this:

sudo /usr/sbin/update-smart-drivedb

then type:

sudo smartctl -i /dev/sda

and post the results.

sudo smartctl -i /dev/sdb
smartctl 5.43 2012-06-30 r3573 [x86_64-linux-3.5.0-26-generic] (local build)
Copyright (C) 2002-12 by Bruce Allen, http://smartmontools.sourceforge.net

=== START OF INFORMATION SECTION ===
Model Family: Seagate Momentus XT (AF)
Device Model: ST750LX003-1AC154
Serial Number: W200GQ2M
LU WWN Device Id: 5 000c50 05c4470af
Firmware Version: SM12
User Capacity: 750,156,374,016 bytes [750 GB]
Sector Sizes: 512 bytes logical, 4096 bytes physical
Device is: In smartctl database [for details use: -P show]
ATA Version is: 8
ATA Standard is: ATA-8-ACS revision 4
Local Time is: Thu Apr 4 19:46:52 2013 CEST
SMART support is: Available - device has SMART capability.
SMART support is: Enabled

[oshunluvr]

This shows 60c as "Operating Temp."

http://www.seagate.com/internal-hard...d-drive/specs/

Seems odd that you don't get the heat using windows though. You can try these mount options in fstab:

noatime,data=writeback,nobh,errors=remount-ro

These will stop time stamping every time a file is accessed, cause metadata to be written during low drive demand times, and to avoid using buffer heads when possible. This will improve performance slightly and thus might let the drive run a bit cooler. The data=writeback and nobh may result in a data corruption if there is a crash during a write operation, so depending on your fault tolerance, you might want to skip those.

What scheduler are you using at the moment?

sudo cat /sys/block/sda/queue/scheduler

Likely, it's CFQ. Try setting it at noop instead:

sudo echo noop > /sys/block/hda/queue/scheduler

[oshunluvr]

Field 190 in your hddtemp printout shows the current temp: 48c

[FrankBarmentlo]

This shows 60c as "Operating Temp."
http://www.seagate.com/internal-hard...d-drive/specs/

Seems odd that you don't get the heat using windows though. You can try these mount options in fstab:

noatime,data=writeback,nobh,errors=remount-ro

These will stop time stamping every time a file is accessed, cause metadata to be written during low drive demand times, and to avoid using buffer heads when possible. This will improve performance slightly and thus might let the drive run a bit cooler. The data=writeback and nobh may result in a data corruption if there is a crash during a write operation, so depending on your fault tolerance, you might want to skip those.

What scheduler are you using at the moment?

sudo cat /sys/block/sda/queue/scheduler

Likely, it's CFQ. Try setting it at noop instead:

sudo echo noop > /sys/block/hda/queue/scheduler

result of ls /sys/block
loop0 loop2 loop4 loop6 ram0 ram10 ram12 ram14 ram2 ram4 ram6 ram8 sda sr0
loop1 loop3 loop5 loop7 ram1 ram11 ram13 ram15 ram3 ram5 ram7 ram9 sdb

your command would be sudo echo noop > /sys/block/sdb/queue/scheduler(sdb is the relevant drive),
but this gives "bash: /sys/block/sdb/queue/scheduler: Permission denied"

[dibl]

your command would be sudo echo noop > /sys/block/sdb/queue/scheduler(sdb is the relevant drive),
but this gives "bash: /sys/block/sdb/queue/scheduler: Permission denied"

Confirmed that the "sudo echo ..." command is not working in my 13.04 Kubuntu system -- something has changed from older versions where that worked to change the scheduler.

If you reboot and press "e" when you see the grub menu, on the linux kernel boot line, you can add the option "scheduler=noop" "elevator=noop" and then continue the boot with Ctrl-X, it should boot with noop enabled. (I didn't have time to test this yet).

[FrankBarmentlo]

thanks for your reply, I am going to try this now,

should i add, or edit the "scheduler=" option?

[kubicle]

Confirmed that the "sudo echo ..." command is not working in my 13.04 Kubuntu system -- something has changed from older versions where that worked to change the scheduler.

This is a common pitfall, you can't use sudo+redirection like that (the redirection is handled by the shell which is not running with elevated privileges), you either have to start a root shell (with 'sudo -i') or do something like:

[B]echo noop | sudo tee /sys/block/hda/queue/scheduler[/B]

[dibl]

Yep, good catch, kubicle (I'm accustomed to working as root in Debian).

OK, so in my previous post I wrote the boot option wrong -- I fixed it -- it is "elevator=noop". If you like the performance of your system better after you reboot and try it, then you can make that the permanent scheduler by editing this line in /etc/default/grub:

GRUB_CMDLINE_LINUX_DEFAULT="quiet elevator=noop"

and then run

sudo update-grub

Another idea would be to slow down the ext4 journalling from the default which is to update the journal every 5 seconds. Of course this has a consequence -- a loss of power could result in a loss of more data with a longer journalling interval. If you wanted to change it to every 20 seconds, for example (4 times less frequently), you could edit the /etc/fstab mount line for /dev/sdb1 (or its UUID) and add the "commit=nn" option. It would then look something like this:

UUID=ec21f5b3-7fd4-4f4b-af8d-cf787b147ae8     /mnt/DATA        ext4         noatime,commit=20   0    0

[FrankBarmentlo]

I will implement this when i am back at linux(playing game in windows right now),
and report back tomorrow around this time.

thanks for it, I really appreciate all help I got here

[SteveRiley]

What scheduler are you using at the moment?

sudo cat /sys/block/sda/queue/scheduler

Likely, it's CFQ. Try setting it at noop instead:

sudo echo noop > /sys/block/hda/queue/scheduler

I would not recommend noop for spinning media. Since noop is just simple FIFO queue, it's likely going to result in much more head movement, thus raising the overall temperature of the drive. CFQ groups I/O requests into batches that minimize seek time, reducing head movement.