linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* 2.4.29: strange behaviour system clock
@ 2005-01-30 19:20 Mario Vanoni
  2005-01-30 22:10 ` Michael Obster
  0 siblings, 1 reply; 2+ messages in thread
From: Mario Vanoni @ 2005-01-30 19:20 UTC (permalink / raw)
  To: linux-kernel

the same with
2.4.29-rc[123]
2.4.28-lck1
2.4.23-aa3
every time repeatable

UP P4-3400HT, 2GB mem, no swap
IDE NEC DVD-RW ND-3500AG (dev/sr0)

DVD with 48 files (*.tar.bz2), 4.4GB

ntpdate -b swisstime.ethz.ch: offset 0.0...
time dircmp /mnt/cdrom /source
thinks it used 20 minutes, no errors
ntpdate -b ...: offset 1134 sec !!!

SMP dual P3-550, 1GB mem, no swap
SCSI PIONER DVD-ROM DVD-304 (slot-in, /dev/sr0)

same DVD, identical source (copied)

ntpdate -b ...: offset 0.0...
time dircmp /mnt/cdrom /source
thinks it used 12 minutes, no errors
ntpdate -b ...: 0.020522 sec

SMP dual Xeon-2800HT, 2GB mem, no swap
IDE NEC DVD-RW ND-2501A (/dev/sr0)

same deviation, offset tons of seconds
this is the production, time must remain correct

Burning CD/DVD with IDE burners,
there not exist SCSI burners,
similar retards of the system clock.
CD using cdrecord, DVD growisofs.

Doing hwclock --show, time is always correct.

Not in LKML, cc if you need, and ... thanks

Mario

PS We burn DVD only since 2 weeks,
   before only CD and only on the SCSI machine,
   and ... NO PROBLEM.


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: 2.4.29: strange behaviour system clock
  2005-01-30 19:20 2.4.29: strange behaviour system clock Mario Vanoni
@ 2005-01-30 22:10 ` Michael Obster
  0 siblings, 0 replies; 2+ messages in thread
From: Michael Obster @ 2005-01-30 22:10 UTC (permalink / raw)
  To: LKML

Hi,

I have also seen such offsets on some boxes running at EADS (a Compaq 
Proliant [<-- this one has a very small offset] and a no name machine 
[<-- this one is terrible in 10 minutes 2-3 minutes too fast]). I have 
no numbers at the moment.
If you need some more information please contact me.

I tried to fix this issue also, but haven't found any solution. Both 
servers are running SuSE Linux I think on 8.x (2.4 kernel). I don't know 
at the moment if we tried already an update to 2.6 kernel.

For the moment our fix is a 10 minute cronjob which execute
hwclock --hctosys
because hardware clock is running correct.

A fix or a hint whats wrong would be very nice :-).

Cheers,
Michael Obster
---
mailto:michael@obster.org
http://www.obster.org

Mario Vanoni wrote:
> the same with
> 2.4.29-rc[123]
> 2.4.28-lck1
> 2.4.23-aa3
> every time repeatable
> 
> UP P4-3400HT, 2GB mem, no swap
> IDE NEC DVD-RW ND-3500AG (dev/sr0)
> 
> DVD with 48 files (*.tar.bz2), 4.4GB
> 
> ntpdate -b swisstime.ethz.ch: offset 0.0...
> time dircmp /mnt/cdrom /source
> thinks it used 20 minutes, no errors
> ntpdate -b ...: offset 1134 sec !!!
> 
> SMP dual P3-550, 1GB mem, no swap
> SCSI PIONER DVD-ROM DVD-304 (slot-in, /dev/sr0)
> 
> same DVD, identical source (copied)
> 
> ntpdate -b ...: offset 0.0...
> time dircmp /mnt/cdrom /source
> thinks it used 12 minutes, no errors
> ntpdate -b ...: 0.020522 sec
> 
> SMP dual Xeon-2800HT, 2GB mem, no swap
> IDE NEC DVD-RW ND-2501A (/dev/sr0)
> 
> same deviation, offset tons of seconds
> this is the production, time must remain correct
> 
> Burning CD/DVD with IDE burners,
> there not exist SCSI burners,
> similar retards of the system clock.
> CD using cdrecord, DVD growisofs.
> 
> Doing hwclock --show, time is always correct.
> 
> Not in LKML, cc if you need, and ... thanks
> 
> Mario
> 
> PS We burn DVD only since 2 weeks,
>   before only CD and only on the SCSI machine,
>   and ... NO PROBLEM.
> 


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2005-01-30 22:10 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2005-01-30 19:20 2.4.29: strange behaviour system clock Mario Vanoni
2005-01-30 22:10 ` Michael Obster

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).