this time with the right email alias.
On Sun, Jan 3, 2016 at 1:01 PM Craig Small wrote:
> Hello Paul,
> I'm chasing up old procps bugs and noticed this one is still opened.
> It's been a long way since a 2.6 kernel and 3.2.x procps, wondering
> if you still get your odd start times?
>
> - Cra
Hello Paul,
I'm chasing up old procps bugs and noticed this one is still opened.
It's been a long way since a 2.6 kernel and 3.2.x procps, wondering
if you still get your odd start times?
- Craig
--
Craig Small (@smallsees) http://enc.com.au/ csmall at : enc.com.au
Debian GNU/Linux
Dear Craig,
> I couldn't see which version of procps you were using.
"dpkg -l" shows "ii procps 1:3.2.8-9 ..." (squeeze up-to-date).
> Is it always just a second earlier? ...
Yes, it seems so. Have not seen any other discrepancies for a long time
now.
> It is looking like a kernel problem, or
On Mon, Jan 30, 2012 at 12:06:27PM +1100, paul.sz...@sydney.edu.au wrote:
> Testing this old thing again at squeeze 6.0.4 with 2.6.32 kernel,
> can still reproduce, the long command in previous message shows:
Hi Paul,
Thanks for letting me know this is still a problem. The only thing is,
I could
Testing this old thing again at squeeze 6.0.4 with 2.6.32 kernel,
can still reproduce, the long command in previous message shows:
$ while :; do \
date; /bin/ps -o lstart,command | grep /ps; date; echo; \
done | \
perl -ne '
BEGIN { $/ = "\n\n" }
@x=grep(s/^(... ... .. ..:..:..).
I can still (2.6.24 kernel, lenny up-to-date) reproduce with command
(wrapped for readability, OK for cut-and-paste into xterm, ctrl-C when
had enough):
while :; do \
date; /bin/ps -o lstart,command | grep /ps; date; echo; \
done | \
perl -ne '
BEGIN { $/ = "\n\n" }
@x=grep(s/^(.
One day after the lenny upgrade and last boot of the machine, there
does not seem to be a great time drift. I was running something like
while :; do
date; /bin/ps -o lstart,command | grep /ps; date
done
and that showed:
...
Thu May 14 13:18:42 EST 2009
Thu May 14 13:18:42 2009 /bin/ps -o lsta
I have now updated my "problem" machine to lenny and 2.6.26 kernel:
r...@pisa:~# cat /etc/debian_version
5.0.1
r...@pisa:~# uname -r
2.6.26-pk03.10-svr
and the problem already is visible two hours after boot:
r...@pisa:~# date; /bin/ps -o lstart,command | grep /ps; date
Wed May 13 11:40:18 EST 2
Dear Craig,
> ... its looking like its not a procps bug. Or
> perhaps there is something procps assumes that is incorrect.
I suspect the latter (this is "idle speculation", without having delved
into the sources).
> ... your process start time is *less* than the uptime ...
> So, the kernel is g
On Mon, Aug 11, 2008 at 12:58:54PM +1000, Paul Szabo wrote:
> Running the commands suggested, I get:
That's interesting, but its looking like its not a procps bug. Or
perhaps there is something procps assumes that is incorrect.
There is no overflow problems here, the manual calculation is agreein
Running the commands suggested, I get:
[EMAIL PROTECTED]:~$ date
Mon Aug 11 12:55:15 EST 2008
[EMAIL PROTECTED]:~$ bash
[EMAIL PROTECTED]:~$ PID=$$; ( date +"%s -" ; cut -f 1 -d ' ' /proc/uptime ;
echo " + (" ; cut -f 22 -d ' ' /proc/$PID/stat ; echo ' /100 )' ) | (tr '\n' '
'; echo ) ; ps -p $P
On Sun, Aug 10, 2008 at 03:13:50PM +1000, Paul Szabo wrote:
> I often observe this, mostly on that one machine "pisa". I do not think
> pisa is "special" in any way: has same hardware as some others, runs the
> same kernel, more-or-less the same daemons. Pisa is still running sarge:
Curious, just a
I still see the problem:
[EMAIL PROTECTED]:~$ date; /bin/ps ux | grep /ps; date
Sun Aug 10 14:45:24 EST 2008
psz 15724 0.0 0.0 2496 844 pts/3R+ 14:38 0:00 /bin/ps ux
psz 15725 0.0 0.0 1548 472 pts/3S+ 14:38 0:00 grep /ps
Sun Aug 10 14:45:24 EST 2008
I often obser
tags 408879 unreproducible
thankyou
On Mon, Feb 26, 2007 at 06:41:24AM -0500, Justin Pryzby wrote:
> Might have expected hz to be accurate to better than .02; on another machine:
Ah, it might be related to the Hertz calculation, but the bug submitters
are not using strange architectures so its not
On Mon, Feb 19, 2007 at 09:50:36AM +1100, Craig Small wrote:
> On Sun, Feb 18, 2007 at 03:38:00PM -0500, Justin Pryzby wrote:
> > I'm of the understanding that ntp uses adjtimex() to do it's stuff.
>
> I suppose the next thing to work out is, is it a kernel problem or a ps
> problem.
I guess (stra
On Sun, Feb 18, 2007 at 03:38:00PM -0500, Justin Pryzby wrote:
> I'm of the understanding that ntp uses adjtimex() to do it's stuff.
I suppose the next thing to work out is, is it a kernel problem or a ps
problem.
- Craig
--
Craig Small GnuPG:1C1B D893 1418 2AF4 45EE 95CB C76C E5AC 12CA
On Mon, Feb 19, 2007 at 06:41:24AM +1100, Paul Szabo wrote:
> Justin,
>
> > ... Do you suspect some particular cause here, too?
>
> Yes, I "blame" ntpd. I only guess that it uses "suspend" to slow
> things down. (No, I wouldn't "manually" suspend my server.)
I'm of the understanding that ntp uses
Justin,
> ... Do you suspect some particular cause here, too?
Yes, I "blame" ntpd. I only guess that it uses "suspend" to slow
things down. (No, I wouldn't "manually" suspend my server.)
Cheers,
Paul Szabo [EMAIL PROTECTED] http://www.maths.usyd.edu.au/u/psz/
School of Mathematics and Stati
On Sun, Feb 18, 2007 at 09:29:05PM +1100, Paul Szabo wrote:
> Justin,
>
> > does this happen to you on a laptop machine, or otherwise?
>
> I do not have Debian laptops. So it is otherwise: on my "main"
> departmental login server
Same for us. You mentioned bug #161633, in which the submitter
Justin,
> does this happen to you on a laptop machine, or otherwise?
I do not have Debian laptops. So it is otherwise: on my "main"
departmental login server
Cheers,
Paul Szabo [EMAIL PROTECTED] http://www.maths.usyd.edu.au/u/psz/
School of Mathematics and Statistics University of Syd
On Mon, Jan 29, 2007 at 09:49:49AM +1100, Paul Szabo wrote:
> Again looking at BTS, this bug seems similar to #161633.
"me too"; Paul, does this happen to you on a laptop machine, or
otherwise?
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL
Again looking at BTS, this bug seems similar to #161633.
Cheers,
Paul Szabo [EMAIL PROTECTED] http://www.maths.usyd.edu.au/u/psz/
School of Mathematics and Statistics University of SydneyAustralia
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble?
Package: procps
Version: 1:3.2.1-2
Severity: normal
File: /bin/ps
We use ntpd to keep time synced. Then somehow the machine uses two times,
a "good" one set by ntpd, and an "internal drifted" one that should never
be shown. Confusingly, ps shows the "wrong" START time:
[EMAIL PROTECTED]:~$ date
23 matches
Mail list logo
