[OT] 1500 days uptime.

[OT] 1500 days uptime.

[Nick Warne]

Hi all,

BrrrrrrrrrrrrBrrrr

That was me blowing my own trumpet again :-)

Re:
http://www.ussg.iu.edu/hypermail/linux/kernel/0407.1/0651.html

Now just hit 1500 days:

-
[nick@486Linux nick]$ last -xf /var/run/utmp runlevel
runlevel (to lvl 3)                    Sun Oct 14 16:07 - 21:41 (1502+06:34)

utmp begins Sun Oct 14 16:07:40 2001
-

Utterly remarkable - the box gets no maintenance at all.

I would love to know how much data it has delivered, but alas, in 2001 I 
wasn't up-to-speed with that sort of thing :-)

Nick
-- 
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb
My quake2 project:
http://sourceforge.net/projects/quake2plus/

Re: [OT] 1500 days uptime.

[Nick Warne]

On Thursday 24 November 2005 22:32, Norbert van Nobelen wrote:
> great uptime. I have 3 readhat boxes who are now more than 1 year up, but
> bizar enough reset the uptime (command uptime) back to zero and started
> counting over again.

Hi Norbert,

Yes, uptime 'wraps' at about 493 days or something.

use:

last -xf /var/run/utmp runlevel

to get last change to run level (run level at boot, presumably).  This was 
posted on LKML ages ago where I learnt to use it to get true 'uptime'.

Nick
-- 
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb
My quake2 project:
http://sourceforge.net/projects/quake2plus/

Re: [OT] 1500 days uptime.

[Nick Warne]

On Thursday 24 November 2005 22:43, Ian McDonald wrote:
>
> Talking of things wrapping, I see the load average counter rolls over at
> 1024. Any hints on getting real load averages when they're above 1024?
>
> --
> ian

1024 load averages? I have never seen anything above 1.5 on any box I run?  
You sure?

Nick
-- 
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb
My quake2 project:
http://sourceforge.net/projects/quake2plus/

Re: [OT] 1500 days uptime.

[Norbert van Nobelen]

load of 1024: Running java with 500 threads? Or just wrote a program which 
just says load=1024?
On Thursday 24 November 2005 23:58, you wrote:
> On Thursday 24 November 2005 22:43, Ian McDonald wrote:
> > Talking of things wrapping, I see the load average counter rolls over at
> > 1024. Any hints on getting real load averages when they're above 1024?
> >
> > --
> > ian
>
> 1024 load averages? I have never seen anything above 1.5 on any box I run?
> You sure?
>
> Nick

-- 
________
www.hipersonik.com : Open source experts

Re: [OT] 1500 days uptime.

[Tim Schmielau]

On Thu, 24 Nov 2005, Nick Warne wrote:
> On Thursday 24 November 2005 22:43, Ian McDonald wrote:
> >
> > Talking of things wrapping, I see the load average counter rolls over at
> > 1024. Any hints on getting real load averages when they're above 1024?

You need to change LOAD_FREQ, FSHIFT, and EXP_1, EXP_5, EXP_15 in
<linux/sched.h>.

  http://sosdg.org/~coywolf/lxr/source/include/linux/sched.h?v=2.6.14#L69

Tim

Re: [OT] 1500 days uptime.

[J. Scott Kasten]

Impressive indeed.  I had a SPARC LX running Red Hat 6.1 or 6.2 I think it 
was, stripped down for a web server and I got 585 days out of it before I 
had to pull the plug to relocate it.  It weathered NIMBDA, an FTP exploit 
and numberous other things for nearly 2 years.

-Scott-

On Thu, 24 Nov 2005, Nick Warne wrote:

> Date: Thu, 24 Nov 2005 21:47:45 +0000
> From: Nick Warne <nick@linicks.net>
> To: linux-kernel@vger.kernel.org
> Subject: [OT] 1500 days uptime.
> 
> Hi all,
>
> BrrrrrrrrrrrrBrrrr
>
> That was me blowing my own trumpet again :-)
>
> Re:
> http://www.ussg.iu.edu/hypermail/linux/kernel/0407.1/0651.html
>
> Now just hit 1500 days:
>
> -
> [nick@486Linux nick]$ last -xf /var/run/utmp runlevel
> runlevel (to lvl 3)                    Sun Oct 14 16:07 - 21:41 (1502+06:34)
>
> utmp begins Sun Oct 14 16:07:40 2001
> -
>
> Utterly remarkable - the box gets no maintenance at all.
>
> I would love to know how much data it has delivered, but alas, in 2001 I
> wasn't up-to-speed with that sort of thing :-)
>
> Nick
> -- 
> "Person who say it cannot be done should not interrupt person doing it."
> -Chinese Proverb
> My quake2 project:
> http://sourceforge.net/projects/quake2plus/
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>

Re: [OT] 1500 days uptime.

[Bill Davidsen]

Nick Warne wrote:
> Hi all,
> 
> BrrrrrrrrrrrrBrrrr
> 
> That was me blowing my own trumpet again :-)
> 
> Re:
> http://www.ussg.iu.edu/hypermail/linux/kernel/0407.1/0651.html
> 
> Now just hit 1500 days:
> 
> -
> [nick@486Linux nick]$ last -xf /var/run/utmp runlevel
> runlevel (to lvl 3)                    Sun Oct 14 16:07 - 21:41 (1502+06:34)
> 
> utmp begins Sun Oct 14 16:07:40 2001
> -
> 
> Utterly remarkable - the box gets no maintenance at all.

But it clearly gets a very reliable flavor of electricity...
> 
> I would love to know how much data it has delivered, but alas, in 2001 I 
> wasn't up-to-speed with that sort of thing :-)

We got one to 1460 or so, then got BSOD on the controller which switches 
from the UPS to the diesel when they get up to speed, dropped power on 
the whole data center (at work).

I ran one at home from the night 1.2.13 was released (or two days after 
when a patch came out) until the afternoon before Y2K, when I decided I 
didn't want to have that system to check at midnight. It was an 8MB 
386SX-16 named "glacial" for its performance rather than because it was 
cool ;-) But it did DNS nicely, which was all I could ask.

I think you have the record, though.
> 
> Nick


-- 
    -bill davidsen (davidsen@tmr.com)
"The secret to procrastination is to put things off until the
  last possible moment - but no longer"  -me

Re: [OT] 1500 days uptime.

[Wakko Warner]

Bill Davidsen wrote:
> Nick Warne wrote:
> >Hi all,
> >
> >BrrrrrrrrrrrrBrrrr
> >
> >That was me blowing my own trumpet again :-)
> >
> >Re:
> >http://www.ussg.iu.edu/hypermail/linux/kernel/0407.1/0651.html
> >
> >Now just hit 1500 days:
> >
> >-
> >[nick@486Linux nick]$ last -xf /var/run/utmp runlevel
> >runlevel (to lvl 3)                    Sun Oct 14 16:07 - 21:41 
> >(1502+06:34)
> >
> >utmp begins Sun Oct 14 16:07:40 2001
> >-
> >
> >Utterly remarkable - the box gets no maintenance at all.
> 
> But it clearly gets a very reliable flavor of electricity...
> >
> >I would love to know how much data it has delivered, but alas, in 2001 I 
> >wasn't up-to-speed with that sort of thing :-)
> 
> We got one to 1460 or so, then got BSOD on the controller which switches 
> from the UPS to the diesel when they get up to speed, dropped power on 
> the whole data center (at work).
> 
> I think you have the record, though.

I'm not sure about that one. =)
I had a Pentium 233MMX pc running for 5 years 3 days.  Booted Nov 18th 1999
and was downed due to a UPS failure around Nov 21 2004.

5 years counting 2 leap year days (2000 and 2004), that's 1827 IIRC.

It was running kernel 2.2.13, 64mb ram, 1.6gb disk, it was a dns server nfs
root server and a NIS server.  Other than that it didn't do much.  It did
get use via the network everyday.  Too bad I don't have a last entry from
utmp before it was killed.

If anyone is interested in the last entry, I'll see if I can dig itup from
old backups.

-- 
 Lab tests show that use of micro$oft causes cancer in lab animals
 Got Gas???

Re: [OT] 1500 days uptime.

[Nick Warne]

On Tuesday 29 November 2005 00:31, Wakko Warner wrote:
> Bill Davidsen wrote:
> > Nick Warne wrote:
> > >Hi all,
> > >
> > >BrrrrrrrrrrrrBrrrr
> > >
> > >That was me blowing my own trumpet again :-)
> > >
> > >Re:
> > >http://www.ussg.iu.edu/hypermail/linux/kernel/0407.1/0651.html
> > >
> > >Now just hit 1500 days:
> > >
> > >-
> > >[nick@486Linux nick]$ last -xf /var/run/utmp runlevel
> > >runlevel (to lvl 3)                    Sun Oct 14 16:07 - 21:41
> > >(1502+06:34)
> > >
> > >utmp begins Sun Oct 14 16:07:40 2001
> > >-
> > >
> > >Utterly remarkable - the box gets no maintenance at all.
> >
> > But it clearly gets a very reliable flavor of electricity...
> >
> > >I would love to know how much data it has delivered, but alas, in 2001 I
> > >wasn't up-to-speed with that sort of thing :-)
> >
> > We got one to 1460 or so, then got BSOD on the controller which switches
> > from the UPS to the diesel when they get up to speed, dropped power on
> > the whole data center (at work).
> >
> > I think you have the record, though.
>
> I'm not sure about that one. =)

No, I wasn't even thinking that - just reporting what a wonderful job it all 
is - and yes, power supply here in Pompey UK is good (but we do pay thru' the 
nose for everything in the UK).  The last time I _did_ reboot that machine 
was when my kettle lead shorted out and blew the fuses to my flats 240v 
supply ring main.

According to the Linux counter site, there are more higher (my machine is 3rd 
in the list):

http://counter.li.org/reports/uptimestats.php

Nick
-- 
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb
My quake2 project:
http://sourceforge.net/projects/quake2plus/

Re: [OT] 1500 days uptime.

[Wakko Warner]

Nick Warne wrote:
> No, I wasn't even thinking that - just reporting what a wonderful job it all 
> is - and yes, power supply here in Pompey UK is good (but we do pay thru' the 
> nose for everything in the UK).  The last time I _did_ reboot that machine 
> was when my kettle lead shorted out and blew the fuses to my flats 240v 
> supply ring main.
> 
> According to the Linux counter site, there are more higher (my machine is 3rd 
> in the list):
> 
> http://counter.li.org/reports/uptimestats.php

I tried to run their scripts on my system, but being as old as it was
(debian hamm I believe), I think I had problems.  Either that or it was
before they knew about the "last -x ..." thing.

I just looked,  Dang, I could have had the record by over 100 days.  Oh
well, maybe my alpha might:
 18:03:07 up 902 days,  1:12,  6 users,  load average: 0.00, 0.00, 0.00
Linux narf 2.4.18 #2 Tue Jun 4 18:32:32 EDT 2002 alpha GNU/Linux

64-bit arch is nice!

-- 
 Lab tests show that use of micro$oft causes cancer in lab animals
 Got Gas???

Re: [OT] 1500 days uptime.

[Nick Warne]

On Tuesday 29 November 2005 23:02, Wakko Warner wrote:
> Nick Warne wrote:

> > http://counter.li.org/reports/uptimestats.php
>
> I tried to run their scripts on my system, but being as old as it was
> (debian hamm I believe), I think I had problems.  Either that or it was
> before they knew about the "last -x ..." thing.

I submitted the 'last -x' patch to them - I think it is in the beta scripts 
now.  I run the 'last -x..' version on the 486 and my gentoo box as I had to 
test it all worked ;-)  I dunno how the other boxes do it > 493 (or whatever 
it is)

> I just looked,  Dang, I could have had the record by over 100 days.  Oh
> well, maybe my alpha might:
>  18:03:07 up 902 days,  1:12,  6 users,  load average: 0.00, 0.00, 0.00
> Linux narf 2.4.18 #2 Tue Jun 4 18:32:32 EDT 2002 alpha GNU/Linux
>
> 64-bit arch is nice!

It's a good site, with a lot of work...  been there a long time too.  Some of 
them kernels I never heard off though.

Nick
-- 
"Person who say it cannot be done should not interrupt person doing it."
-Chinese Proverb
My quake2 project:
http://sourceforge.net/projects/quake2plus/

Re: [OT] 1500 days uptime.

[Wakko Warner]

Nick Warne wrote:
> I submitted the 'last -x' patch to them - I think it is in the beta scripts 
> now.  I run the 'last -x..' version on the 486 and my gentoo box as I had to 
> test it all worked ;-)  I dunno how the other boxes do it > 493 (or whatever 
> it is)

497 days.  (2^32 / 8640000 days IIRC)

-- 
 Lab tests show that use of micro$oft causes cancer in lab animals
 Got Gas???

Re: [OT] 1500 days uptime.

[jdow]

From: "Nick Warne" <nick@linicks.net>

> On Tuesday 29 November 2005 00:31, Wakko Warner wrote:
>> Bill Davidsen wrote:
>> > Nick Warne wrote:
>> > >Hi all,
>> > >
>> > >BrrrrrrrrrrrrBrrrr
>> > >
>> > >That was me blowing my own trumpet again :-)
>> > >
>> > >Re:
>> > >http://www.ussg.iu.edu/hypermail/linux/kernel/0407.1/0651.html
>> > >
>> > >Now just hit 1500 days:
>> > >
>> > >-
>> > >[nick@486Linux nick]$ last -xf /var/run/utmp runlevel
>> > >runlevel (to lvl 3)                    Sun Oct 14 16:07 - 21:41
>> > >(1502+06:34)
>> > >
>> > >utmp begins Sun Oct 14 16:07:40 2001
>> > >-
>> > >
>> > >Utterly remarkable - the box gets no maintenance at all.
>> >
>> > But it clearly gets a very reliable flavor of electricity...
>> >
>> > >I would love to know how much data it has delivered, but alas, in 2001 I
>> > >wasn't up-to-speed with that sort of thing :-)
>> >
>> > We got one to 1460 or so, then got BSOD on the controller which switches
>> > from the UPS to the diesel when they get up to speed, dropped power on
>> > the whole data center (at work).
>> >
>> > I think you have the record, though.
>>
>> I'm not sure about that one. =)
> 
> No, I wasn't even thinking that - just reporting what a wonderful job it all 
> is - and yes, power supply here in Pompey UK is good (but we do pay thru' the 
> nose for everything in the UK).  The last time I _did_ reboot that machine 
> was when my kettle lead shorted out and blew the fuses to my flats 240v 
> supply ring main.
> 
> According to the Linux counter site, there are more higher (my machine is 3rd 
> in the list):
> 
> http://counter.li.org/reports/uptimestats.php

There is an interesting point here that is worth noting. I don't think I
would BEGIN to try this with a 2.6 based kernel. And I am not sure it is
doable yet with a 2.4 kernel if the machine is to be exposed to the wild.

{o.o}    Joanne Dow

Re: [OT] 1500 days uptime.

[Willy Tarreau]

On Tue, Nov 29, 2005 at 03:08:10PM -0800, jdow wrote:
> From: "Nick Warne" <nick@linicks.net>
(...)
> >No, I wasn't even thinking that - just reporting what a wonderful job it 
> >all is - and yes, power supply here in Pompey UK is good (but we do pay 
> >thru' the nose for everything in the UK).  The last time I _did_ reboot 
> >that machine was when my kettle lead shorted out and blew the fuses to 
> >my flats 240v supply ring main.
> >
> >According to the Linux counter site, there are more higher (my machine 
> >is 3rd in the list):
> >
> >http://counter.li.org/reports/uptimestats.php
> 
> There is an interesting point here that is worth noting. I don't think I
> would BEGIN to try this with a 2.6 based kernel. And I am not sure it is
> doable yet with a 2.4 kernel if the machine is to be exposed to the wild.

it may depend on what you do with it. I reached 1000 days with a
2.4.19pre5 plus a lot of patches a few years ago. However, it would
have been dangerous to leave this machine exposed because there were
so many local root exploits that the smallest vulnerability in any
installed service would have cause a disaster. But I also have some
machines running 2.4.22+some patches which have not been rebooted
since kernel upgrade, and which are used as front firewalls on a
big fat internet pipe, and they are still up and running normally.
So I can say that 2.4 was stable enough 2 years ago, and is even
more right now. Judging from the many people reporting years of
uptime (and from local experience), it seems that 2.4 reached such
a stability around 2.4.18.

> {o.o}    Joanne Dow

Regards,
Willy

Re: [OT] 1500 days uptime.

[Bill Davidsen]

jdow wrote:

> From: "Nick Warne" <nick@linicks.net>
>
>> On Tuesday 29 November 2005 00:31, Wakko Warner wrote:
>>
>>> Bill Davidsen wrote:
>>> > Nick Warne wrote:
>>> > >Hi all,
>>> > >
>>> > >BrrrrrrrrrrrrBrrrr
>>> > >
>>> > >That was me blowing my own trumpet again :-)
>>> > >
>>> > >Re:
>>> > >http://www.ussg.iu.edu/hypermail/linux/kernel/0407.1/0651.html
>>> > >
>>> > >Now just hit 1500 days:
>>> > >
>>> > >-
>>> > >[nick@486Linux nick]$ last -xf /var/run/utmp runlevel
>>> > >runlevel (to lvl 3)                    Sun Oct 14 16:07 - 21:41
>>> > >(1502+06:34)
>>> > >
>>> > >utmp begins Sun Oct 14 16:07:40 2001
>>> > >-
>>> > >
>>> > >Utterly remarkable - the box gets no maintenance at all.
>>> >
>>> > But it clearly gets a very reliable flavor of electricity...
>>> >
>>> > >I would love to know how much data it has delivered, but alas, in 
>>> 2001 I
>>> > >wasn't up-to-speed with that sort of thing :-)
>>> >
>>> > We got one to 1460 or so, then got BSOD on the controller which 
>>> switches
>>> > from the UPS to the diesel when they get up to speed, dropped 
>>> power on
>>> > the whole data center (at work).
>>> >
>>> > I think you have the record, though.
>>>
>>> I'm not sure about that one. =)
>>
>>
>> No, I wasn't even thinking that - just reporting what a wonderful job 
>> it all is - and yes, power supply here in Pompey UK is good (but we 
>> do pay thru' the nose for everything in the UK).  The last time I 
>> _did_ reboot that machine was when my kettle lead shorted out and 
>> blew the fuses to my flats 240v supply ring main.
>>
>> According to the Linux counter site, there are more higher (my 
>> machine is 3rd in the list):
>>
>> http://counter.li.org/reports/uptimestats.php
>
>
> There is an interesting point here that is worth noting. I don't think I
> would BEGIN to try this with a 2.6 based kernel. And I am not sure it is
> doable yet with a 2.4 kernel if the machine is to be exposed to the wild.
>
> {o.o}    Joanne Dow
>
I got 0.99.4 to stay up 16 weeks ;-)

And the old GE/Honeywell 600 crashed due to rollover at 35 days, and we 
hit the bug first three years after it was coded. Hardware is better!

-- 
bill davidsen <davidsen@tmr.com>
  CTO TMR Associates, Inc
  Doing interesting things with small computers since 1979