Re: [LUG] uptime oddity

["Theo P. Zourzouvillys" <theo@xxxxxxxxxxxxxxxx>]

On Tuesday 27 January 2004 00:29, Neil Williams wrote:
> ?Well, it turns out that in a 32-bit number, you can store 497.1 days
> before the number wraps.

indeed, and then oddd things start happening when they do, pre 2.4.19.  from 
one of my servers:

# uptime
00:49:46 up 99 days,  7:09,  1 user,  load average: 0.00, 0.00, 0.00

# ps aux | grep 2005
root     20372  0.0  0.0  2748 1156 ?        S     2005   0:00 /usr/sbin/sshd

(note it appears to have been started in 2005)...

# date
Wed Jan 28 00:50:42 GMT 2004

it's actually been up for 497 + 99 days.  the kernel just completly looses the 
plot when it comes to processes starting dates.

on another mahcine:

# uptime
Unknown HZ value! (43) Assume 100.
01:00:55 up 439 days,  8:49,  1 user,  load average: 0.03, 0.12, 0.15

Which iirc is a bug in 2.4.18 with regards to jiffie rollovers.

Didn't we cover this a while ago? ;)

~ Theo

-- 
Theo Zourzouvllys
Operations Manager
Associated Networks (UK) Limited
- theo@xxxxxxxx
- 0207 1270 501

--
The Mailing List for the Devon & Cornwall LUG
Mail majordomo@xxxxxxxxxxxx with "unsubscribe list" in the
message body to unsubscribe.