[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Reducing HDD writing affect on whole system.

On Sat, 15 Oct 2011 22:02:45 +0700, Sthu Deus wrote:

>>Consider running a smartctl test on the disk, it can be dying or having
>>a severe hardware problem.
> I have run this way:
> smartctl --test=short /dev/sda
> in 2 minutes it completed and in its log (-l selftest /dev/sda) I saw
> its report stating that everything is OK that is no errors found. Seems
> to me it makes not speed tests but whither an HDD alive, I tried to take
> long test but it seems to do the same thing just hard way and not
> performance test...
> Or I have done something wrong? - I'm new to the utility though.

Smartctl (best if you run the "long" test) scans for errors on the 
filesystem structure, but cannot tell about other physical errors (bad 
cable or connector problems), so yes, better do additional hard disk 
speed tests.

>>I think you should first find out what makes this (freezes and slowness)
>>happens because if you're facing a hardware related issue, there will be
>>no improvements in tweaking the filesystem mount options, sooner or
>>later your disk will fail and you will only delay its agony.
> Well it is my long experience (about 7 years already) w/ SATA HDDs that
> to me it works slower (on Linux) than the PATA ones - I understand it
> sounds very weird, but so it is. 

Yes, indeed it sounds weird :-)

> Long time ago I have seen for example -
> the SATA HDD w/ less capacity was fsck-ed much longer than PATA one
>   w/ having more capacity - of course there was not heard of / used ext4
>   at all, ext3 - was the highest version of the FS. And so on. Till this
>   very day I suppose the culprit is the hardware drivers in the kernel
>   or wherever - if it is not so - I mean that a lot of people have high
>   writing/reading speeds of SATAs then I have something to do w/ my
>   configurations :) though I do not have a glue.
> Having said that I'm not looking for finding a bottle neck w/ the discs,
> but rather reduce its effect that it has on my systems - at least this
> is the point that I had when I started the thread.

On standard systems setups (where hdds are connected directly to the sata 
controller of the board) and non-complex layouts (non-raid/lvm/
clusters...) I've never experienced freezes nor delays with sata nor ide 
hard disks. Not "visible" delays, I mean.

Of course, on older/less powerful systems I can't get the same 
performance than I have on modern hardware (faster hard disks, faster 
sata contoller, more ram, faster cpu...) but neither freezes nor 
slowdowns on normal setups unless you are intentionally putting to the 
limit your hardware.

So, what kind of applications are you running on your system and over 
what hardware?

> Yet if I'm wrong, and the devices can produce high performance in Linux
> / Debian, then I would find the problem rather than lowing writing
> process priority, etc. In this case would You help to investigate the
> problem? I supply any info relating the matter.

After discarding a hardware problem, you can run hdd stress tests to find 
out what can be the cause of the delays you are having. There are also 
dedicated LiveCD suites aimed to benchmark computers at higher loads.



Reply to: