[ale] uptime?

[Grant Anderson]

My turn.


I think that uptime is important.  When was the last time your car died on you?

Uptime was important to you then I would suspect.  Otherwise you take if for

granted.


Also, if you are used to your car being in the shop every month, then it's not

unusual for you.  It would be unusual then, however, to go a year without having

to take the car in.


In my experience, and in the cumulative experience others have reported to me,

Windows boxes have more problems with memory leaks and thus need to be rebooted

more often than Unix boxes.  One can always find exceptions to this but this does

not change the reality.  One also does not see NT/2000 boxes replacing high end

Unix boxes in enterprise server situations.  (Maybe one day...maybe not.)


One can argue back and forth, swap "exceptional" tales and experiences, and kick

the subject around or even to death.  What I see out in the field is that NT/2000

boxes have to be rebooted much more often than Unix and Linux boxes.  This speaks

favorably for Unix/Linux in this regard.  


>From a personal experience standpoint I regularly have had throughout the last

couple of years have had to reboot my NT/2000 workstations and servers and I'm 

quite tired of it.  I wish Microsoft would fix this.  Anyone that has an NT box

up and running for months or years is, in my opinion, lucky and should be happy

with such fortuitous good fortune.  


I also think that all current operating systems should be more robust, have more

uptime, and have less need for rebooting.  Linux is very good in this regard but

there is room for improvement I'm sure.


There I did it....stated my opinions....fanned the flames of THE UPTIME WAR and 

the MICROSOFT - NT WAR....and who knows what else?!


Compute in Peace,


   Grant




-----Original Message-----

From: Luis Luna [mailto:luis at btr-architects.com]
To: ale at ale.org

Sent: Thursday, February 22, 2001 10:31 AM

To: Atlanta Linux Enthusiasts

Subject: RE: [ale] uptime?



        I believe that, uptime is great when you have a good streak going. But I

had two nt 4.0 server stay up for 1 year 4 months without rebooting or

hiccuping. I finally shut them down, pulled out the hard drives / cards /

etc. and slapped them on a new mobo, PIII chip, RAM, and case. Re-applied

the service pack, rebooted, and they have been running since 12/15/2000 like

champs. So what. I built a RH 6.2 server at the same time, loaded samba and

had it share files out along with the 2 NT boxes, I haven't needed to shut

it down for the same time as the NT boxes.

        I have an NT box acting as our proxy server and other internet jobs, P200

with 96 megs of RAM, it has been up for 7 months. I only had to reboot

because of some software I loaded requested a reboot, so before that it had

been running for the previous 8 months since I built it.


luis


-----Original Message-----

F


> The point is, extended uptimes speaks of stability and robust kernel and

> application code.  One of the problems I believe NT has is memory

> leaks.  I used to have to run NT on my laptop, and it would just plain

> go brain dead after a while.


In the interest of getting a good flamewar started...


I actually have 2 NT4 SP4 boxen running WINS/DNS (internal only) that

have somehow managed to stay up since 04/21/2000. In contrast, a Samba

server running on identical hardware only lasted 4 months. After some

analysis on the box in question, I believe I've found a faulty power

supply feeding "dirty juice" to the Linux box. To make a long story

short...


Uptimes don't mean sh*t!!!


I've seen HP-UX machines run for a year at a time, only to crash HARD 2

mos after the next reboot. I have friends in the military who run

NT4/Exchange 5.5 boxes that run for 6-8 months at a time without so much

as a hiccup. I've also seen Linux machines run for over a year. But I've

also seen Linux machines crash and burn. There are so many factors

involved in maintaining uptime that it's not worth attempting to compare

OS statistics.


I'll have to say that it's been my experience that Linux/BSD systems do

seem to run longer. I've never personally seen a BSD box crash. However, I

think the hardware, system load, and configuration have much more to do

with it than the OS. There are plenty of things for us (Linux enthusiasts)

to brag about without resorting to using uptime statistics.


--

Jonathan Rickman

X Corps Security

http://www.xcorps.net



--

To unsubscribe: mail majordomo at ale.org with "unsubscribe ale" in message

body.



--

To unsubscribe: mail majordomo at ale.org with "unsubscribe ale" in message body.