LinuxQuestions.org
Support LQ: Use code LQ3 and save $3 on Domain Registration
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Slackware
User Name
Password
Slackware This Forum is for the discussion of Slackware Linux.

Notices


Reply
  Search this Thread
Old 01-06-2014, 04:39 AM   #31
roderich
LQ Newbie
 
Registered: Dec 2013
Posts: 15

Original Poster
Rep: Reputation: Disabled

Quote:
Originally Posted by Richard Cranium View Post
What happens when you power-cycle versus reboot?

One of your experiments may have set something on the drive that persists until power cycle. I only ask because a reboot isn't exactly the same thing as a power cycle.
Well, there *may* be more hidden secrets. The case has arrived at a point where you begin to think about the influence of moon phases or stray alpha particles.

Seriously: For the moment I am in peace again with my PC and not in the mood for more restart exercises. When I see the system's health degrading again I can of course try a next restart with a power-cycle for a change (so far I had made only soft reboots).
 
Old 01-06-2014, 04:58 AM   #32
roderich
LQ Newbie
 
Registered: Dec 2013
Posts: 15

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by guanx View Post
Any mechanical resonance?

Loose SATA cable?

Failing disk (check with smartctl)?

Or (though unlikely from your previous posts) a degraded RAID being rebuilt?
No RAID involved, it's just a single disk.
And per recommendation by someone else here I had already run a big 4-hours internal disk selftest without errors.

My main suspect is still some system (memory?) corruption, perhaps in connection with the discussed PAE "features". Another indication for this is that I had also unusual problems with Firefox in the same time: YouTube videos did not play any longer and it crashed completely when I shut it down before shutdown of the system.
 
Old 01-06-2014, 09:43 AM   #33
Richard Cranium
Senior Member
 
Registered: Apr 2009
Location: Carrollton, Texas
Distribution: Slackware64 14.1
Posts: 1,724

Rep: Reputation: 541Reputation: 541Reputation: 541Reputation: 541Reputation: 541Reputation: 541
Quote:
Originally Posted by roderich View Post
Seriously: For the moment I am in peace again with my PC and not in the mood for more restart exercises. When I see the system's health degrading again I can of course try a next restart with a power-cycle for a change (so far I had made only soft reboots).
No sweat. Most of us buy our computers for reasons other than using them as debugging exercisers.
 
Old 01-06-2014, 11:04 AM   #34
roderich
LQ Newbie
 
Registered: Dec 2013
Posts: 15

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by Richard Cranium View Post
No sweat. Most of us buy our computers for reasons other than using them as debugging exercisers.
Well said ....

While it has been my job and hobby for many years to diagnose and repair computer malfunctions, I am meanwhile much more content when they "just work".
 
Old 01-06-2014, 12:15 PM   #35
perbh
Member
 
Registered: May 2008
Location: Republic of Texas
Posts: 275

Rep: Reputation: 48
I have a rig somewhat similar to yours but 'longer in the tooth', 8 gigs memory (using PAE) and a 1TB disk (stock 14.1)
Code:
# hdparm -i /dev/sda

/dev/sda:

 Model=WDC WD2500SD-01KCC0, FwRev=08.02D08, SerialNo=WD-WCAL74231724
 Config={ HardSect NotMFM HdSw>15uSec SpinMotCtl Fixed DTR>5Mbs FmtGapReq }
 RawCHS=16383/16/63, TrkSize=57600, SectSize=600, ECCbytes=74
 BuffType=DualPortCache, BuffSize=8192kB, MaxMultSect=16, MultSect=16
 CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=488397168
 IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
 PIO modes:  pio0 pio1 pio2 pio3 pio4 
 DMA modes:  mdma0 mdma1 mdma2 
 UDMA modes: udma0 udma1 udma2 udma3 udma4 udma5 *udma6 
 AdvancedPM=no WriteCache=enabled
 Drive conforms to: Unspecified:  ATA/ATAPI-1,2,3,4,5,6
Then
Code:
# hdparm -tT /dev/sda

/dev/sda:
 Timing cached reads:   3478 MB in  2.00 seconds = 1740.56 MB/sec
 Timing buffered disk reads: 172 MB in  3.01 seconds =  57.07 MB/sec
and finally
Code:
# dd if=/dev/zero of=/disk/tf bs=4M count=1000 conv=fdatasync #jfs in this case
1000+0 records in
1000+0 records out
4194304000 bytes (4.2 GB) copied, 85.4185 s, 49.1 MB/s
Not the sharpest knife in the drawer - but it will do.

Last edited by perbh; 01-06-2014 at 12:19 PM.
 
Old 06-04-2015, 08:42 PM   #36
kik.lozev
LQ Newbie
 
Registered: Feb 2010
Posts: 3

Rep: Reputation: 1
Seems like I am not the only one with this problem (Yey!)

I have 2 PCs both running stock 32bit slackware 14.1 and they both have similar hardware:
sda -> solid state disk for the OS and sdb -> 1TB data for /home; both ext4 filesystems.
Both of them have 16Gigs of ram.
My laptop is asus with intel i7 and my desktop is with latest (as of writing) AMD processor.

I observe the exact same issue on both of them - in random intervals the io of the sdb becomes unbearably slow ~5mb/s and a lot of disk intensive processes fall in IO wait state and become unresponsive.
I noticed that after hibernation on the laptop this issue is there, but on my desktop it presents itself randomly - sometimes it runs for weeks without a trace of slow disks or whatsoever.

For some reason when I disabled the disk barriers of the sdb the laptop seems to be running fine after the sleep/wake and the issue seems to be gone (at least for now).

I will definitely look into PAE but since this issue is not consistent at all its very difficult to debug.
 
Old 08-20-2015, 05:04 AM   #37
kik.lozev
LQ Newbie
 
Registered: Feb 2010
Posts: 3

Rep: Reputation: 1
I hound a solution for my problem.

Short story:
Code:
echo 1 > /proc/sys/vm/highmem_is_dirtyable
and putting it in the /etc/rc.d/rc.local solved the issue.


Long story:
Turns out its all about how Linux manages the "dirty memory" - portions of the RAM that are used to buffer/cache the disk writes before flushing them to the disk.
(I am not deeply familiar with the way kernel manages the memory so I will be happy to see others' replies about this.)
In the /proc/sys/vm/ you will find ways to tune the way kernel manages the dirty memory. Read those for more info:Most users will want to tune "dirty_background_ratio" and "dirty_ratio" - those percentages control the amount of virtual memory that is used for buffering disk writes, but here comes the issue - since I am using a 32bit kernel with PAE extension kernel prefers not to use the high memory for caching since its slower - this is why "highmem_is_dirtyable" is 0 by default. And since I have 16gigs of RAM those percentages create a great difference. So when a program attempts to write large amounts of data to the disk it starts at extremely high speed and almost immediately drops to several kbps after it reaches the dirty pages limits.

It might be a kernel bug when calculating the "dirty_ratio" based on the total amount of memory and not taking into account "highmem_is_dirtyable" but again - I am not familiar enough to fire kernel bugs
 
Old 08-20-2015, 07:01 AM   #38
syg00
LQ Veteran
 
Registered: Aug 2003
Location: Australia
Distribution: Lots ...
Posts: 13,349

Rep: Reputation: 1303Reputation: 1303Reputation: 1303Reputation: 1303Reputation: 1303Reputation: 1303Reputation: 1303Reputation: 1303Reputation: 1303Reputation: 1303
Have a read of this - particularly the 32-bit vs 64-bit issues.
Linus is very unlikely to accept bugs against 32-bit with large RAM - he has ranted against this years ago.
 
1 members found this post helpful.
Old 08-21-2015, 08:41 PM   #39
metaschima
Senior Member
 
Registered: Dec 2013
Distribution: Slackware
Posts: 1,898

Rep: Reputation: 476Reputation: 476Reputation: 476Reputation: 476Reputation: 476
16 GB RAM and 32-bit ... I agree with Linus, he should not fix any bugs for this configuration because it is nonsensical.
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
How can external disk be mounted so a ordinary user can write to the disk Glenn D. Linux - Newbie 1 10-06-2012 02:32 AM
Sata disk write speed is very slow! eltonsky Debian 9 09-16-2010 06:07 AM
Extremely slow write on disk with high CPU usage aliojjati Linux - Newbie 1 09-04-2010 01:29 AM
slow disk reads: Slackware 10 kernel 2.6.7 nirvanix Slackware 7 06-29-2004 11:26 AM
slackware: slow disk access ? Menestrel Slackware 3 05-22-2004 05:46 PM


All times are GMT -5. The time now is 02:50 PM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration