Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.os.linux.hardware > #3043 > unrolled thread

HD writes hog RAM with SATA HD replacing an IDE

Back to article view | Back to comp.os.linux.hardware

Contents

  HD writes hog RAM with SATA HD replacing an IDE Mike Spencer <mds@bogus.nodomain.nowhere> - 2016-04-11 02:47 -0300
    Re: HD writes hog RAM with SATA HD replacing an IDE philo <philo@privacy.net> - 2016-04-11 04:19 -0500
      Re: HD writes hog RAM with SATA HD replacing an IDE Mike Spencer <mds@bogus.nodomain.nowhere> - 2016-04-13 00:06 -0300
        Re: HD writes hog RAM with SATA HD replacing an IDE Charlie Gibbs <cgibbs@kltpzyxm.invalid> - 2016-04-13 17:56 +0000
        Re: HD writes hog RAM with SATA HD replacing an IDE philo <philo@privacy.net> - 2016-04-13 13:26 -0500
    Re: HD writes hog RAM with SATA HD replacing an IDE "Carlos E.R." <robin_listas@invalid.es> - 2016-04-11 11:30 +0200

#3043 — HD writes hog RAM with SATA HD replacing an IDE

I've had to replace a failed 80G IDE drive with a 150G SATA drive.

Writes to the new HD seem to be hogging memory, stalling X response
and other HD reads such as causing  music playing from a .wav file on
HD to stutter.

After reinstall of Slackware 11 [1] from original media, recovery of
files from backups and tweaking all the idiosyncratic configs, I'm
back to just where I was before the failure except for this bogging
down on big HD writes.

These, which write the HD, cause the problem:

    cp USB-stick/bigfile  ->  hda3
    cp hda3/bigfile       ->  hda4

but these, which only read the HD, don't:

    cp hda3/bigfile         -> USB-stick
    cp USB-stick/bigfile    -> Other USB-stick
    mplayer hda3/some-video.mp4

I have P4 CPU, Slack 11, 2.4.33 kernel and 256M of RAM with which this
problem didn't occur before replacing the HD.

The new SATA HD plugs into one of two ports on the IBM mobo labeled
"SATA IDE 1" and "SATA IDE 2".   The drive is seen by Linux as hda,
not sda. (Some on-mobo emulation I don't know about?)

I have /sys but it's empty.

/proc/sys/vm contains the following files:

    bdflush
    block_dump
    kswapd
    laptop_mode
    max-readahead
    max_map_count
    min-readahead
    overcommit_memory
    page-cluster
    pagetable_cache
    vm_anon_lru
    vm_cache_scan_ratio
    vm_gfp_debug
    vm_lru_balance_ratio
    vm_mapped_ratio
    vm_passes
    vm_vfs_scan_ratio



It seems like I should be able to fix this problem -- apparently vm
hogging RAM on HD writes -- with a tweak to one (or more) of those
files.

After looking at /usr/src/linux-2.4.33.3/Documentation/sysctl/vm.txt,
I've tried (one at a time):

       Reducing max-readahead      from  31  to  15
       Reducing bdflush [interval] from 500  to 100
       Reducing bdflush [sync]     from  60  to  20

with no effect.

Any specific suggestions on fixing this?

- Mike


[1] Yes, there's a  reason for using Slack 11, not relevant here.

-- 
Mike Spencer                  Nova Scotia, Canada

[toc] | [next] | [standalone]

#3044

On 04/11/2016 12:47 AM, Mike Spencer wrote:
> I've had to replace a failed 80G IDE drive with a 150G SATA drive.
>
> Writes to the new HD seem to be hogging memory, stalling X response
> and other HD reads such as causing  music playing from a .wav file on
> HD to stutter.
>
> After reinstall of Slackware 11 [1] from original media, recovery of
> files from backups and tweaking all the idiosyncratic configs, I'm
> back to just where I was before the failure except for this bogging
> down on big HD writes.
>
> These, which write the HD, cause the problem:
>
>      cp USB-stick/bigfile  ->  hda3
>      cp hda3/bigfile       ->  hda4
>


Looks like a defective hard drive, run the manufacturer's diagnostic.

Due to the small size of the drive I'm guessing that when you said "new" 
you simply meant "new to the machine" and not an actual new drive?

[toc] | [prev] | [next] | [standalone]

#3048

philo <philo@privacy.net> writes:

> Looks like a defective hard drive, run the manufacturer's diagnostic.

The problematic bigfile -> HD copies proceed quickly.  It appears to
be a memory usage problem during writes.

> Due to the small size of the drive I'm guessing that when you said "new" 
> you simply meant "new to the machine" and not an actual new drive?

Right.  A used 150G. I have a new 500G SATA coming.  But I'm not above
reverting to an 80G IDE drive if t he prblem becomes annoying enough.

From Stephenson's The System of the World, Saturn (Peter Hoxton) &
Daniel Waterhouse in Hockley in the Hole:

    "Mechanic since I was a lad, clock-maker since I came to my
    senses," said Saturn.  "The piece of information you are wanting,
    Doctor, is this: this here is an old Hooke balance-spring watch,
    this is. When the Master made it, why, it might've been the best
    time-piece ever fashioned by human hands. But now there's a score
    of proper horologists round Clerkenwell who can make ones that'll
    keep better time. Technology ages, dunnit"?

    Daniel pursed his lips to keep from laughing at the spectacle of
    this new, five-guinea word, Technology, emerging from that head.

    "It ages faster'n we do. It can be difficult for a bloke to keep
    up."

    "Is that your story, Saturn? You could not keep up, and so you
    went bad"?

    "I grew weary of keeping up, Doctor. That is my story, if you must
    know. I grew weary of transitory knowledge, and decided to seek
    knowledge of a more aeternal nature."

-- 
Mike Spencer                  Nova Scotia, Canada

[toc] | [prev] | [next] | [standalone]

#3049

On 2016-04-13, Mike Spencer <mds@bogus.nodomain.nowhere> wrote:

> From Stephenson's The System of the World, Saturn (Peter Hoxton) &
> Daniel Waterhouse in Hockley in the Hole:
>
>     "Mechanic since I was a lad, clock-maker since I came to my
>     senses," said Saturn.  "The piece of information you are wanting,
>     Doctor, is this: this here is an old Hooke balance-spring watch,
>     this is. When the Master made it, why, it might've been the best
>     time-piece ever fashioned by human hands. But now there's a score
>     of proper horologists round Clerkenwell who can make ones that'll
>     keep better time. Technology ages, dunnit"?
>
>     Daniel pursed his lips to keep from laughing at the spectacle of
>     this new, five-guinea word, Technology, emerging from that head.
>
>     "It ages faster'n we do. It can be difficult for a bloke to keep
>     up."
>
>     "Is that your story, Saturn? You could not keep up, and so you
>     went bad"?
>
>     "I grew weary of keeping up, Doctor. That is my story, if you must
>     know. I grew weary of transitory knowledge, and decided to seek
>     knowledge of a more aeternal nature."

Beautiful.  Or, as I like to put it:

    Some people want to be timely.  I'd rather be timeless.

-- 
/~\  cgibbs@kltpzyxm.invalid (Charlie Gibbs)
\ /  I'm really at ac.dekanfrus if you read it the right way.
 X   Top-posted messages will probably be ignored.  See RFC1855.
/ \  HTML will DEFINITELY be ignored.  Join the ASCII ribbon campaign!

[toc] | [prev] | [next] | [standalone]

#3050

On 04/12/2016 10:06 PM, Mike Spencer wrote:
> philo <philo@privacy.net> writes:
>
>> Looks like a defective hard drive, run the manufacturer's diagnostic.
>
> The problematic bigfile -> HD copies proceed quickly.  It appears to
> be a memory usage problem during writes.
>


OK, it was just a guess on my part.



>> Due to the small size of the drive I'm guessing that when you said "new"
>> you simply meant "new to the machine" and not an actual new drive?
>
> Right.  A used 150G. I have a new 500G SATA coming.  But I'm not above
> reverting to an 80G IDE drive if t he prblem becomes annoying enough.


GOOD
  V      (just realized there is ^ > <  but no down arrow)
>
>  From Stephenson's The System of the World, Saturn (Peter Hoxton) &
> Daniel Waterhouse in Hockley in the Hole:
>
>      "Mechanic since I was a lad, clock-maker since I came to my
>      senses," said Saturn.  "The piece of information you are wanting,
>      Doctor, is this: this here is an old Hooke balance-spring watch,
>      this is. When the Master made it, why, it might've been the best
>      time-piece ever fashioned by human hands. But now there's a score
>      of proper horologists round Clerkenwell who can make ones that'll
>      keep better time. Technology ages, dunnit"?
>
>      Daniel pursed his lips to keep from laughing at the spectacle of
>      this new, five-guinea word, Technology, emerging from that head.
>
>      "It ages faster'n we do. It can be difficult for a bloke to keep
>      up."
>
>      "Is that your story, Saturn? You could not keep up, and so you
>      went bad"?
>
>      "I grew weary of keeping up, Doctor. That is my story, if you must
>      know. I grew weary of transitory knowledge, and decided to seek
>      knowledge of a more aeternal nature."
>

[toc] | [prev] | [next] | [standalone]

#3045

On 2016-04-11 07:47, Mike Spencer wrote:
> I've had to replace a failed 80G IDE drive with a 150G SATA drive.
> 
> Writes to the new HD seem to be hogging memory, stalling X response
> and other HD reads such as causing  music playing from a .wav file on
> HD to stutter.

...

> These, which write the HD, cause the problem:
> 
>     cp USB-stick/bigfile  ->  hda3
>     cp hda3/bigfile       ->  hda4

hda3? Not sda3? Wow.


> I have P4 CPU, Slack 11, 2.4.33 kernel and 256M of RAM with which this
> problem didn't occur before replacing the HD.

2.4.33 kernel? That's quite old, might explain your problem. 2007.

> The new SATA HD plugs into one of two ports on the IBM mobo labeled
> "SATA IDE 1" and "SATA IDE 2".   The drive is seen by Linux as hda,
> not sda. (Some on-mobo emulation I don't know about?)

Or this.

Even my IDE disks are seen as /dev/sdX in all my systems, never
/dev/hdX. But I have not seen a 2.4 kernel in years, except in embedded
machines.

-- 
Cheers, Carlos.


--- news://freenews.netfront.net/ - complaints: news@netfront.net ---

[toc] | [prev] | [standalone]

Back to top | Article view | comp.os.linux.hardware

csiph-web