Yeah, I think he's a little ... off.
Brian Kelsay
djgoku <> 03/22/05 10:32AM >>>
On Mon, 18 Apr 2005 19:22:06 -0500, Oren Beck <hotmail.com> wrote:
Q- what REALLY killed Frank Herbert?
A- Seeing the David Lynching of Dune.
Oren
" Did Omnius run on Linux" ?
Date: Mon, 18 Apr 2005 19:22:06 -0500 <------A little off? From: Oren Beck <@hotmail.com> User-Agent: Mozilla Thunderbird 0.8 (Windows/20040913)
Oren, you might want to update your clock and your Thunderbird installation.. =)
he's from the future, wondering why our clocks are a month behind.
On Tue, 22 Mar 2005 10:48:27 -0600, Brian Kelsay [email protected] wrote:
Yeah, I think he's a little ... off.
Brian Kelsay
djgoku <> 03/22/05 10:32AM >>>
On Mon, 18 Apr 2005 19:22:06 -0500, Oren Beck <hotmail.com> wrote:
It seems that someone hasn't discovered ntp ...
On Tue, 22 Mar 2005, David Nicol wrote:
he's from the future, wondering why our clocks are a month behind.
On Tue, 22 Mar 2005 10:48:27 -0600, Brian Kelsay [email protected] wrote:
Yeah, I think he's a little ... off.
Brian Kelsay
djgoku <> 03/22/05 10:32AM >>>
On Mon, 18 Apr 2005 19:22:06 -0500, Oren Beck <hotmail.com> wrote:
Kclug mailing list [email protected] http://kclug.org/mailman/listinfo/kclug
//========================================================\ || D. Hageman [email protected] || \========================================================//
On Tue, March 22, 2005 1:47 pm, D. Hageman said:
It seems that someone hasn't discovered ntp ...
Ntp doesn't guarantee correct time, it's just another way to screw it up. (Like gentoo overwriting the config with garbage by default.)
Jonathan Hutchins wrote:
On Tue, March 22, 2005 1:47 pm, D. Hageman said:
It seems that someone hasn't discovered ntp ...
Ntp doesn't guarantee correct time, it's just another way to screw it up. (Like gentoo overwriting the config with garbage by default.)
It _does_ guarantee that you have the same time as a server or a set of peers, within a given tolerance. This is helpful if you want to correlate log data on different hosts, or run "make" in an NFS-mounted directory.
On Tue, March 22, 2005 4:02 pm, Gerald Combs said:
Jonathan Hutchins wrote:
Ntp doesn't guarantee correct time, it's just another way to screw it up.
It _does_ guarantee that you have the same time as a server or a set of peers, within a given tolerance. This is helpful if you want to correlate log data on different hosts, or run "make" in an NFS-mounted directory.
That's if/when it's configured and working correctly. Then it'll even provide the correct date for your email timestamps. That's not necessarily a given though.
I had a system that, because I was relying on ntp, and because ntp was confirured correctly when I checked after setting it up, was six hours off after subsequent reboots. Gentoo had, in the absence of useful data from the DHCP server, overwritten the valid ntp config file with default garbage. (This is the default behavior for gentoo.) Caveat admin.
On Tue, 22 Mar 2005, Jonathan Hutchins wrote:
On Tue, March 22, 2005 4:02 pm, Gerald Combs said:
Jonathan Hutchins wrote:
Ntp doesn't guarantee correct time, it's just another way to screw it up.
It _does_ guarantee that you have the same time as a server or a set of peers, within a given tolerance. This is helpful if you want to correlate log data on different hosts, or run "make" in an NFS-mounted directory.
That's if/when it's configured and working correctly. Then it'll even provide the correct date for your email timestamps. That's not necessarily a given though.
I had a system that, because I was relying on ntp, and because ntp was confirured correctly when I checked after setting it up, was six hours off after subsequent reboots. Gentoo had, in the absence of useful data from the DHCP server, overwritten the valid ntp config file with default garbage. (This is the default behavior for gentoo.) Caveat admin.
*cough* Works well on RedHat systems since like ... 6.0 *cough*
:-)
//========================================================\ || D. Hageman [email protected] || \========================================================//
Strange, I never had any problems with ntp when I was running gentoo. Setting it up in debian was a no brainer, and I never have any problems with my time being off by very much. Ntp is a great tool to have, but it's missing one thing: a plug-in for displaying Aztec dates and times.
Brian D.
--- Jonathan Hutchins wrote:
On Tue, March 22, 2005 4:02 pm, Gerald Combs said:
Jonathan Hutchins wrote:
Ntp doesn't guarantee correct time, it's just
another way to screw it
up.
It _does_ guarantee that you have the same time as
a server or a set of
peers, within a given tolerance. This is helpful
if you want to
correlate log data on different hosts, or run
"make" in an NFS-mounted
directory.
That's if/when it's configured and working correctly. Then it'll even provide the correct date for your email timestamps. That's not necessarily a given though.
I had a system that ... was six hours off after subsequent reboots. Gentoo had, in the absence of useful data from the DHCP server, overwritten the valid ntp config file with default garbage. (This is the default behavior for gentoo.) Caveat admin.
__________________________________ Do you Yahoo!? Read only the mail you want - Yahoo! Mail SpamGuard. http://promotions.yahoo.com/new_mail
I like ddate personally.
Brian, you should code up an adate and submit it to the util-linux people. ;-)
At any rate - I didn't want this e-mail to be completely fluff, so here is some advice.
Non-technical:
Some motherboards have a seriously bad system clock on them. NTP works by slowly skewing the time to keep it in-line with the servers it is feeding from. If your motherboards clock is so bad that NTP believes it can't keep up by tiny skews ... then it will give up and stop working. My recommendation for you in this case is have a nightly cron job that stops ntpd and calls ntpdate and then starts ntpd again. You should schedule the job at a time when it will least effect the system (i.e. without a lot of users or cron jobs dependent on time). In generally doing it on a daily basis prevents the skew from being bad enough for it to cause any real issues doing this. My experience is that even the worst system clocks will take at least a couple of weeks to get so out of wack before ntp will give up.
Technical:
The units of time it works in is called a 'tick' and the actual system call used to slowly augment the time is 'adjtimex'. A smaller unit of time exists as well called a 'jiffie'.
On Tue, 22 Mar 2005, Jack wrote:
Strange, I never had any problems with ntp when I was running gentoo. Setting it up in debian was a no brainer, and I never have any problems with my time being off by very much. Ntp is a great tool to have, but it's missing one thing: a plug-in for displaying Aztec dates and times.
Brian D.
--- Jonathan Hutchins wrote:
On Tue, March 22, 2005 4:02 pm, Gerald Combs said:
Jonathan Hutchins wrote:
Ntp doesn't guarantee correct time, it's just
another way to screw it
up.
It _does_ guarantee that you have the same time as
a server or a set of
peers, within a given tolerance. This is helpful
if you want to
correlate log data on different hosts, or run
"make" in an NFS-mounted
directory.
That's if/when it's configured and working correctly. Then it'll even provide the correct date for your email timestamps. That's not necessarily a given though.
I had a system that ... was six hours off after subsequent reboots. Gentoo had, in the absence of useful data from the DHCP server, overwritten the valid ntp config file with default garbage. (This is the default behavior for gentoo.) Caveat admin.
Do you Yahoo!? Read only the mail you want - Yahoo! Mail SpamGuard. http://promotions.yahoo.com/new_mail _______________________________________________ Kclug mailing list [email protected] http://kclug.org/mailman/listinfo/kclug
//========================================================\ || D. Hageman [email protected] || \========================================================//
On Wed, Mar 23, 2005 at 01:36:32AM -0600, D. Hageman wrote:
Some motherboards have a seriously bad system clock on them.
Yes very true.
Also I've noticed that the time will be very off if the BIOS battery is dying and the machine is powered off. I don't know if that's actually true or something "the network is slow when it rains" type thing.
For your Googling pleasure:
http://www.beaglesoft.com/mainfaqclock.htm http://www.smartcomputing.com/editorial/article.asp?article=articles/archive...
Jeremy
Also I've noticed that the time will be very off if the BIOS battery is dying and the machine is powered off.
At startup after network is up I run rdate -s 10.0.0.1 (10.0.0.1 is my local ntpd server, running OpenNTPD) after that OpenNTP client keeps my time up-to-date with local ntpd server.
jo
On Tuesday 22 March 2005 10:51 pm, Jack wrote:
Strange, I never had any problems with ntp ...
Ever notice how people who completely lack experience in problems or troubleshooting of a certain thing, application or whatever, proclaim it as if their ignorance were some kind of virtue?
Hi,
On Tue, Mar 22, 2005 at 05:37:33PM -0600, Jonathan Hutchins wrote:
On Tue, March 22, 2005 4:02 pm, Gerald Combs said:
Jonathan Hutchins wrote:
Ntp doesn't guarantee correct time, it's just another way to screw it up.
It _does_ guarantee that you have the same time as a server or a set of peers, within a given tolerance. This is helpful if you want to correlate log data on different hosts, or run "make" in an NFS-mounted directory.
That's if/when it's configured and working correctly. Then it'll even provide the correct date for your email timestamps. That's not necessarily a given though.
I had a system that, because I was relying on ntp, and because ntp was confirured correctly when I checked after setting it up, was six hours off after subsequent reboots. Gentoo had, in the absence of useful data from the DHCP server, overwritten the valid ntp config file with default garbage. (This is the default behavior for gentoo.) Caveat admin.
Six hours is like "forever" to ntpd. It doesn't like to make big changes to the system clock and will creep up to the correct time a few ms. at a time. You should call "hwclock" to set the hardware clock time to the system time when you shutdown the machine because the system time is set from the hardware clock at boot time. I can't remember if I set this up or if it was done for me by RedHat. There is also an "adjtimex" package that will keep track of the drift of your hardware clock and the last time it was set to try to set the system clock correctly. You can also go into the startup scripts for ntpd and add a call to either "rdate" or "ntpdate" to get the clock set to something reasonable.
A clock that is six hours off sounds like it could be a problem with a Linux system that likes to keep the hardware clock set to UTC and "some other OS" that likes to keep the hardware clock set to local time on a dual boot box. If this is the case you can solve the problem by using "fdisk" to change the type of the partition with "that other OS" to type 82 (linux swap).
Somewhat OT, does anybody know if the Dave Mills of ntp is the same Dave Mills that worked for Motorola in KC back around the time of the 6800/6809? -- Jim
Jonathan Hutchins wrote:
I had a system that, because I was relying on ntp, and because ntp was confirured correctly when I checked after setting it up, was six hours off after subsequent reboots. Gentoo had, in the absence of useful data from the DHCP server, overwritten the valid ntp config file with default garbage. (This is the default behavior for gentoo.) Caveat admin.
Nagios automagically sends me an email whenever NTP is wonky on one of my machines. I thought this happened to everyone.
On Wednesday 23 March 2005 10:01 am, Gerald Combs wrote:
Nagios automagically sends me an email whenever NTP is wonky on one of my machines. I thought this happened to everyone.
What kind of load does Nagios place on a machine? Most of the systems I want to monitor are somewhat marginal - which is why they need monitoring - and on at least one of them external monitoring might not be possible.
Jonathan Hutchins wrote:
What kind of load does Nagios place on a machine? Most of the systems I want to monitor are somewhat marginal - which is why they need monitoring - and on at least one of them external monitoring might not be possible.
The amount of load that Nagios incurs depends on the number of "service" entries you have defined. We have ~100 services defined for ~20 hosts, and the load is minimal, currently wavering between 0.08 and 0.15 on our server (a P3/1000).
On Wednesday 23 March 2005 11:41 am, Gerald Combs wrote:
Jonathan Hutchins wrote:
What kind of load does Nagios place on a machine? Most of the systems I want to monitor are somewhat marginal - which is why they need monitoring
- and on at least one of them external monitoring might not be possible.
The amount of load that Nagios incurs depends on the number of "service" entries you have defined. We have ~100 services defined for ~20 hosts, and the load is minimal, currently wavering between 0.08 and 0.15 on our server (a P3/1000).
How about RAM usage?
Jonathan Hutchins wrote:
On Wednesday 23 March 2005 11:41 am, Gerald Combs wrote:
The amount of load that Nagios incurs depends on the number of "service" entries you have defined. We have ~100 services defined for ~20 hosts, and the load is minimal, currently wavering between 0.08 and 0.15 on our server (a P3/1000).
How about RAM usage?
"top" says 40 MB virtual, 8.7 MB resident.
Well, time is a very relative thing . . . Who is to say what is correct?
I personally like to keep my system clocks sync'ed with the computers that are connected to atomic clocks as they seem to be generally accepted as ... correct. :-)
On Tue, 22 Mar 2005, Jonathan Hutchins wrote:
On Tue, March 22, 2005 1:47 pm, D. Hageman said:
It seems that someone hasn't discovered ntp ...
Ntp doesn't guarantee correct time, it's just another way to screw it up. (Like gentoo overwriting the config with garbage by default.)
Kclug mailing list [email protected] http://kclug.org/mailman/listinfo/kclug
//========================================================\ || D. Hageman [email protected] || \========================================================//
Really? Or shall we consider that no one posting to a Lug list needs NTP!
A dedicated army of Open Source time observers will happily inform you of an out of sync condition.
From: David Nicol [email protected] Reply-To: David Nicol [email protected] To: Brian Kelsay [email protected] CC: [email protected] Subject: Re: Dune Parody skit. Was RE: whitewigs Date: Tue, 22 Mar 2005 12:32:44 -0600 MIME-Version: 1.0 Received: from kclug.org ([139.146.133.42]) by mc11-f36.hotmail.com with Microsoft SMTPSVC(6.0.3790.211); Tue, 22 Mar 2005 10:33:09 -0800 Received: from localhost (localhost.localdomain [127.0.0.1])by kclug.org (Postfix) with ESMTP id A86D5113052;Tue, 22 Mar 2005 12:31:15 -0600 (CST) Received: from kclug.org ([127.0.0.1])by localhost (kclug [127.0.0.1]) (amavisd-new, port 10024) with ESMTPid 06308-03; Tue, 22 Mar 2005 12:31:15 -0600 (CST) Received: from kclug.org (localhost.localdomain [127.0.0.1])by kclug.org (Postfix) with ESMTP id DF2AD14ECC1;Tue, 22 Mar 2005 12:31:05 -0600 (CST) Received: from localhost (localhost.localdomain [127.0.0.1])by kclug.org (Postfix) with ESMTP id 573D114ECBDfor [email protected]; Tue, 22 Mar 2005 12:31:03 -0600 (CST) Received: from kclug.org ([127.0.0.1])by localhost (kclug [127.0.0.1]) (amavisd-new, port 10024) with ESMTPid 06076-10 for [email protected];Tue, 22 Mar 2005 12:30:55 -0600 (CST) Received: from rproxy.gmail.com (rproxy.gmail.com [64.233.170.195])by kclug.org (Postfix) with ESMTP id 1698814ECBCfor [email protected]; Tue, 22 Mar 2005 12:30:54 -0600 (CST) Received: by rproxy.gmail.com with SMTP id a41so1488760rngfor [email protected]; Tue, 22 Mar 2005 10:32:44 -0800 (PST) Received: by 10.38.90.20 with SMTP id n20mr5772686rnb;Tue, 22 Mar 2005 10:32:44 -0800 (PST) Received: by 10.38.22.43 with HTTP; Tue, 22 Mar 2005 10:32:44 -0800 (PST) X-Message-Info: vAu4ZEtdRijM27yeJGcnL2xF/rAvtPEurNqyJdgqaas= X-Original-To: [email protected] Delivered-To: [email protected] DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com;h=received:message-id:date:from:reply-to:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:references;b=sThEC6I4LNZ0t9J81psVC7HiJ2JKTL9SaN1R4lduys/Du4tKRcrX6gaTc3nwyH+5EKqVj65FSWFT4DQIxXUyFaIZhzWM6g8CIz+HNw7qtwxaCv//dISuwYy3hfIspw6H9Ex67tPxdZfKiaESEYRDTKGP77w8uUI9SD9tWa9aiG4= References: [email protected] X-Virus-Scanned: by amavisd-new-20030616-p10 (Debian) at kclug.org X-BeenThere: [email protected] X-Mailman-Version: 2.1.5 Precedence: list List-Id: KCLUG mailing list <kclug.kclug.org> List-Unsubscribe: http://kclug.org/mailman/listinfo/kclug,mailto:[email protected]?subject=unsubscribe List-Archive: http://kclug.org/pipermail/kclug List-Post: mailto:[email protected] List-Help: mailto:[email protected]?subject=help List-Subscribe: http://kclug.org/mailman/listinfo/kclug,mailto:[email protected]?subject=subscribe Errors-To: [email protected] X-Virus-Scanned: by amavisd-new-20030616-p10 (Debian) at kclug.org Return-Path: [email protected] X-OriginalArrivalTime: 22 Mar 2005 18:33:10.0093 (UTC) FILETIME=[992DE7D0:01C52F0D]
he's from the future, wondering why our clocks are a month behind.
On Tue, 22 Mar 2005 10:48:27 -0600, Brian Kelsay [email protected] wrote:
Yeah, I think he's a little ... off.
Brian Kelsay
djgoku <> 03/22/05 10:32AM >>>
On Mon, 18 Apr 2005 19:22:06 -0500, Oren Beck <hotmail.com> wrote:
Kclug mailing list [email protected] http://kclug.org/mailman/listinfo/kclug
-
Brian Kelsay -
D. Hageman -
David Nicol -
djgoku -
Gerald Combs -
Jack -
Jeremy Turner -
Jonathan Hutchins -
Oren Beck -
Uncle Jim