Constant hdd writing/reading? Crazy stuff!

Svip · 06-06-2006, 09:27 AM

I can constantly hear my hdd being (ab)used. I am wondering what I can do to limit this. Because my system is very slow when running X, well... slow at starting up applications, when they are running, it's almost like flying.

I just wish to know reasons for a lot of hdd activity, when I had ubuntu on this very machine, the hdd activity was much less active.

KeithE · 06-06-2006, 09:48 AM

Quote:

Originally Posted by Svip

I can constantly hear my hdd being (ab)used. I am wondering what I can do to limit this. Because my system is very slow when running X, well... slow at starting up applications, when they are running, it's almost like flying.

I just wish to know reasons for a lot of hdd activity, when I had ubuntu on this very machine, the hdd activity was much less active.

How much swap space did you allocate on your hard drive? Or maybe you have a memory leak somewhere (Older version of Firefox? Mplayer has had a few issues on my machine as well) that's eating up the swap space you did set up. Try logging out and back in again and see if that fixes the problem.

Svip · 06-06-2006, 09:56 AM

Quote:

Originally Posted by KeithE

How much swap space did you allocate on your hard drive? Or maybe you have a memory leak somewhere (Older version of Firefox? Mplayer has had a few issues on my machine as well) that's eating up the swap space you did set up. Try logging out and back in again and see if that fixes the problem.

Firefox is newest version. I am not using MPlayer... right now. I have allocated 512MB for swap. Is that stupid? Swap is currently at 0.50, and as I said, running applications rarely have an problem.

davidsrsb · 06-06-2006, 10:03 AM

What does top show on memory usage?
It's also worth running hdparm to check your disk throughput is ok.

Svip · 06-06-2006, 10:09 AM

Quote:

Originally Posted by davidsrsb

What does top show on memory usage?
It's also worth running hdparm to check your disk throughput is ok.

Memory is at 1.00, yes.

hdparm for hda1 and hda4:

Code:

/dev/hda1:
 multcount    = 16 (on)
 IO_support   =  0 (default 16-bit)
 unmaskirq    =  0 (off)
 using_dma    =  1 (on)
 keepsettings =  0 (off)
 readonly     =  0 (off)
 readahead    =  8 (on)
 geometry     = 9733/255/63, sectors = 152248005, start = 4112640

Code:

/dev/hda4:
 multcount    = 16 (on)
 IO_support   =  0 (default 16-bit)
 unmaskirq    =  0 (off)
 using_dma    =  1 (on)
 keepsettings =  0 (off)
 readonly     =  0 (off)
 readahead    =  8 (on)
 geometry     = 9733/255/63, sectors = 4112577, start = 63

tubatodd · 06-06-2006, 12:32 PM

This may NOT have to do with your hard drive however I recently worked on a friends computer and the hard drive they had came with S.M.A.R.T. and SMART detected a problem with the hard drive. It was doing the very same thing (while running Windows). Hard drive access was slow, but programs running in RAM were fine. According to the users guide from the hard drive....the drive was toast. I installed a new one and everything was fixed. I HOPE it isn't a hardware problem, but those symptoms sounded familiar.

kodon · 06-06-2006, 01:10 PM

what wm/de do you use?

after an upgrade a long time ago, my hdd light
would be on constantly under kde. switched to a wm
...problem solved.

antis · 06-06-2006, 02:42 PM

If you are using KDE try disabling the KDED Media Manager.
Have a look at one of my previous posts here: http://www.linuxquestions.org/questi...d.php?t=386967

Svip · 06-06-2006, 03:07 PM

I am using dropline GNOME, for your information.

@tubatodd: I certainly don't hope so.

drkstr · 06-06-2006, 03:41 PM

I am curious about this as well. I noticed a lot that my computer would just start cranking away at the hard drive when I wasn't even using it. I don't seem to be having the problem anymore since I upgraded my memory/cpu, but I would still like to see what my computer is doing when it thinks I am asleep

Does anyone have any recommendations on how to spy on what's going on in my system after it goes idle?

thanks!
...drkstr

gilead · 06-06-2006, 05:17 PM

drkstr, have you looked at the sysstat package from http://perso.orange.fr/sebastien.god...mentation.html? I think it does what you're looking for.

drkstr · 06-06-2006, 05:24 PM

Quote:

drkstr, have you looked at the sysstat ackage from http://perso.orange.fr/sebastien.god...mentation.html?

No I haven't, and yes, that is exactly what I am looking for. Thanks for the link!

...drkstr

Old_Fogie · 06-06-2006, 10:59 PM

I do notice that at midnight the linux goes crazy on every pc. I think that's the indexing tho.

kodon · 06-06-2006, 11:27 PM

Quote:

Originally Posted by Old_Fogie

I do notice that at midnight the linux goes crazy on every pc. I think that's the indexing tho.

the default is 4:40am

theoffset · 06-07-2006, 12:45 AM

Just to eliminate the chance, why don't you run a SMART test on the disk?

Of course, you need a SMART capable BIOS and HD.

Here's what you have to do**:
(**First than anything, you may want to light-read "man smartctl" and "man smartd" -I said light-read because these man pages are huge)

Enter the BIOS (reboot the machine and press the magic key) and enable SMART (some BIOSes enable it for all the Hard Drives, some others just for selected HDs).
Re-boot into Linux and login as root

If you had been running /usr/sbin/smartd (And had SMART enabled in the BIOS) then you can check the system logs (in a default Slackware syslog config all the smartd messages goes to /var/log/messages) for any sign of errors.
Run
Code:
```
smartctl -c </dev/hda or your hard disk name>
```
Check the value of "Extended self test routine recommended polling time".

Now run the actual self test:

Code:

smartctl -t long </dev/hda or your hard disk device name>

You should wait the number of minutes recommended in the "Extended self test routine recommended polling time".

Then check the results

Code:

smartctl -a </dev/hda or your hard disk device name>

You can get just the overall result with

Code:

smartctl -H </dev/hda or your hard disk device name>

Most of what is stated when running "smartctl -a" is pretty self explanatory, except maybe the "Vendor specific SMART attributes with Thresholds", basically if any attribute with a TYPE of Pre_fail has its VALUE lower than THRESH, then your HD is about to collapse.

Hopefully it will not be a HD issue.

Good Luck.