• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Uptime Bragging rights :)

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Originally posted by: n0cmonkey
Originally posted by: jliechty
Originally posted by: n0cmonkey
The load average numbers
give the number of jobs in the run queue averaged over 1, 5 and 15 min-
utes.
Help a little bit?
Yes, tremendously. Thanks. 🙂

Thank the OpenBSD man page writers. Another example of how OpenBSD's docs are top notch 😉

Kewl 😉 Since it's the number of jobs in queue, why is it not an integer value? Most of the time you'll see something like 0.52, 1.21, x.xx?
 
Originally posted by: ugh
Originally posted by: n0cmonkey
Originally posted by: jliechty
Originally posted by: n0cmonkey
The load average numbers
give the number of jobs in the run queue averaged over 1, 5 and 15 min-
utes.
Help a little bit?
Yes, tremendously. Thanks. 🙂

Thank the OpenBSD man page writers. Another example of how OpenBSD's docs are top notch 😉

Kewl 😉 Since it's the number of jobs in queue, why is it not an integer value? Most of the time you'll see something like 0.52, 1.21, x.xx?

I think it has to do with the run, wait, sleep stuff. Cant give you a good answer off hand though.
 
Getting XP to be up for 2 weeks i think is no problem barring any hardware problems that one can have.

I have had mine up for 2 weeks 4days. then wanted to install something so had to reboot.

actually by the time i was done i was like what the heck just lets reboot for the heck of it.

BTW. are u guys counting standby times also..cause when i did this..my computer was on standby at nights...which i dont think should do anything to the stability of the system. if it working well then it should be work fine other times as well..
 
Originally posted by: sak
Getting XP to be up for 2 weeks i think is no problem barring any hardware problems that one can have.

I have had mine up for 2 weeks 4days. then wanted to install something so had to reboot.

actually by the time i was done i was like what the heck just lets reboot for the heck of it.

BTW. are u guys counting standby times also..cause when i did this..my computer was on standby at nights...which i dont think should do anything to the stability of the system. if it working well then it should be work fine other times as well..

My laptop was generally working a DC project when I wasnt using it. I dont think there is a way to take standby time out of the uptime command, so I cant say for sure how long mine was on standby/sleep.
 
first of all, it very much depends what people are running on their server. If it is just DNS, DHCP and/or SMTP server, even Windows 3.11 will run for years. You need to realize that 486 will be more than enough to run those, while in today's world most people use PIII and above for those simple tasks. I mean my Cisco PIX 501 firewall has 33MHz processor with 16MB of RAM performing above functions and it has been up for 3 years now 🙂

However, when you run intense applications such as SQL server, .NET applictions server, etc... and do some sort of productive work on your computer, it is normal for computer to misbehave once in a while. After all, there are other uses for computer besides sitting idle 80.00% of the time and being proud that is has been there doing nothing. There is nothing there to crash it 🙂
 
Originally posted by: n0cmonkey
Originally posted by: ugh
Originally posted by: n0cmonkey
Originally posted by: jliechty
Originally posted by: n0cmonkey
The load average numbers
give the number of jobs in the run queue averaged over 1, 5 and 15 min-
utes.
Help a little bit?
Yes, tremendously. Thanks. 🙂

Thank the OpenBSD man page writers. Another example of how OpenBSD's docs are top notch 😉

Kewl 😉 Since it's the number of jobs in queue, why is it not an integer value? Most of the time you'll see something like 0.52, 1.21, x.xx?

I think it has to do with the run, wait, sleep stuff. Cant give you a good answer off hand though.

It seems to be sort of averaged out. If you have a few runaway processes, and your load is at 2,3,5, whatever, and then kill them all, and just keep running "uptime" over and over, once a second or so, you'll see the first number slowly drop over a period of time, instead of changing instantly, and then eventually it arrives at whatever your current load really is.
 
Originally posted by: andrey
first of all, it very much depends what people are running on their server. If it is just DNS, DHCP and/or SMTP server, even Windows 3.11 will run for years. You need to realize that 486 will be more than enough to run those, while in today's world most people use PIII and above for those simple tasks. I mean my Cisco PIX 501 firewall has 33MHz processor with 16MB of RAM performing above functions and it has been up for 3 years now 🙂

However, when you run intense applications such as SQL server, .NET applictions server, etc... and do some sort of productive work on your computer, it is normal for computer to misbehave once in a while. After all, there are other uses for computer besides sitting idle 80.00% of the time and being proud that is has been there doing nothing. There is nothing there to crash it 🙂

true however my linux box at home is my firewall, proxy, webserver with many sql backends it has been up for over 2 months and is running at 100#5 cpu 24/7 crushing seti units if it was going to crash i think it would have done long before now so short of power outages i cant see a reason for it to die. admittledy its got 2.5gig or ram init so not much is swapped but still impressive none the less cant see win 3.1 running 100% cpu for 2 months flat out can you 🙂
 
However, when you run intense applications such as SQL server, .NET applictions server, etc... and do some sort of productive work on your computer, it is normal for computer to misbehave once in a while

I can understand your workstation having problems, because you're fiddling with things all the time, but unless you're breaking your SQL server frequently it should run flawlessly no matter how much CPU time it consumes constantly. The fact that it's doing a lot of calculations and/or I/O does not make it more prone to breakage, well atleast not with decent software =)
 
My company has a very strange policy in maintaining their production servers. They reboot ALL the servers once a week. These are all Sun servers we're talking about. Apparently this is "good policy"... Umm... 😕
 
It all comes down to the bane of windows operating systems: memory leaks. I beleive that is the major reason that the average windows server or desktop is not able to keep up with Unices and Netware OS. The inflexability if specifized configurations and lack of control are also factors.

Why have the software installed for file and print sharing, a gui and a webbrowser, etc etc, when all you are doing is providing a dns and ftp service or a specialized database server? I know unused stuff doesn't have a effect, but they are still a potential source of confict. What if you do a quick look-up on some odd problem on the internet and a weird java app slowly begins a runaway in the background? come back 2 days later and whala no more memory left. what is the instictive thing for a windows user to do? reboot... What is the instictive thing for a unix user to do? find the offending proccess and kill it.

Not that that means a whole lot. 5 min downtime to reboot vs 10 extra minutes of degragaded performance, but hey. It will defenatly effect uptime.

Thats enough for windows people to say that it is easy POSSIBLE to run w2k for months with no reboot.

And why it is NORMAL for unix people to expect months of reliable uptime.

And as for Netware... Netware is just a specialized peice of rock-solid software. (however i do find it ironic that you need a dos partition to start it up, though hehe)
 
Originally posted by: BingBongWongFooey
Originally posted by: n0cmonkey
Originally posted by: ugh
Originally posted by: n0cmonkey
Originally posted by: jliechty
Originally posted by: n0cmonkey
The load average numbers
give the number of jobs in the run queue averaged over 1, 5 and 15 min-
utes.
Help a little bit?
Yes, tremendously. Thanks. 🙂

Thank the OpenBSD man page writers. Another example of how OpenBSD's docs are top notch 😉

Kewl 😉 Since it's the number of jobs in queue, why is it not an integer value? Most of the time you'll see something like 0.52, 1.21, x.xx?

I think it has to do with the run, wait, sleep stuff. Cant give you a good answer off hand though.

It seems to be sort of averaged out. If you have a few runaway processes, and your load is at 2,3,5, whatever, and then kill them all, and just keep running "uptime" over and over, once a second or so, you'll see the first number slowly drop over a period of time, instead of changing instantly, and then eventually it arrives at whatever your current load really is.

I just wanted to point out that sometimes quotes can get out of hand.
 
Originally posted by: Smilin
Originally posted by: BingBongWongFooey
Originally posted by: n0cmonkey
Originally posted by: ugh
Originally posted by: n0cmonkey
Originally posted by: jliechty
Originally posted by: n0cmonkey
The load average numbers
give the number of jobs in the run queue averaged over 1, 5 and 15 min-
utes.
Help a little bit?
Yes, tremendously. Thanks. 🙂

Thank the OpenBSD man page writers. Another example of how OpenBSD's docs are top notch 😉

Kewl 😉 Since it's the number of jobs in queue, why is it not an integer value? Most of the time you'll see something like 0.52, 1.21, x.xx?

I think it has to do with the run, wait, sleep stuff. Cant give you a good answer off hand though.

It seems to be sort of averaged out. If you have a few runaway processes, and your load is at 2,3,5, whatever, and then kill them all, and just keep running "uptime" over and over, once a second or so, you'll see the first number slowly drop over a period of time, instead of changing instantly, and then eventually it arrives at whatever your current load really is.

I just wanted to point out that sometimes quotes can get out of hand.


Really?

 
Originally posted by: Smilin
Originally posted by: Smilin
Originally posted by: BingBongWongFooey
Originally posted by: n0cmonkey
Originally posted by: ugh
Originally posted by: n0cmonkey
Originally posted by: jliechty
Originally posted by: n0cmonkey
The load average numbers
give the number of jobs in the run queue averaged over 1, 5 and 15 min-
utes.
Help a little bit?
Yes, tremendously. Thanks. 🙂

Thank the OpenBSD man page writers. Another example of how OpenBSD's docs are top notch 😉

Kewl 😉 Since it's the number of jobs in queue, why is it not an integer value? Most of the time you'll see something like 0.52, 1.21, x.xx?

I think it has to do with the run, wait, sleep stuff. Cant give you a good answer off hand though.

It seems to be sort of averaged out. If you have a few runaway processes, and your load is at 2,3,5, whatever, and then kill them all, and just keep running "uptime" over and over, once a second or so, you'll see the first number slowly drop over a period of time, instead of changing instantly, and then eventually it arrives at whatever your current load really is.

I just wanted to point out that sometimes quotes can get out of hand.


Really?

Yeah.


 
Originally posted by: Smilin
Originally posted by: Smilin
Originally posted by: Smilin
Originally posted by: BingBongWongFooey
Originally posted by: n0cmonkey
Originally posted by: ugh
Originally posted by: n0cmonkey
Originally posted by: jliechty
Originally posted by: n0cmonkey
The load average numbers
give the number of jobs in the run queue averaged over 1, 5 and 15 min-
utes.
Help a little bit?
Yes, tremendously. Thanks. 🙂

Thank the OpenBSD man page writers. Another example of how OpenBSD's docs are top notch 😉

Kewl 😉 Since it's the number of jobs in queue, why is it not an integer value? Most of the time you'll see something like 0.52, 1.21, x.xx?

I think it has to do with the run, wait, sleep stuff. Cant give you a good answer off hand though.

It seems to be sort of averaged out. If you have a few runaway processes, and your load is at 2,3,5, whatever, and then kill them all, and just keep running "uptime" over and over, once a second or so, you'll see the first number slowly drop over a period of time, instead of changing instantly, and then eventually it arrives at whatever your current load really is.

I just wanted to point out that sometimes quotes can get out of hand.


Really?

Yeah.

ROFLMAO 😀

 
Originally posted by: Smilin
Originally posted by: Smilin
Originally posted by: Smilin
Originally posted by: BingBongWongFooey
Originally posted by: n0cmonkey
Originally posted by: ugh
Originally posted by: n0cmonkey
Originally posted by: jliechty
Originally posted by: n0cmonkey
The load average numbers
give the number of jobs in the run queue averaged over 1, 5 and 15 min-
utes.
Help a little bit?
Yes, tremendously. Thanks. 🙂
Thank the OpenBSD man page writers. Another example of how OpenBSD's docs are top notch 😉

Kewl 😉 Since it's the number of jobs in queue, why is it not an integer value? Most of the time you'll see something like 0.52, 1.21, x.xx?
I think it has to do with the run, wait, sleep stuff. Cant give you a good answer off hand though.
It seems to be sort of averaged out. If you have a few runaway processes, and your load is at 2,3,5, whatever, and then kill them all, and just keep running "uptime" over and over, once a second or so, you'll see the first number slowly drop over a period of time, instead of changing instantly, and then eventually it arrives at whatever your current load really is.
I just wanted to point out that sometimes quotes can get out of hand.
Really?
Yeah.
Woah, this is getting a little out of hand. Anyway, my thanks go to all those (n0c, bbwf, etc.) who clarified this issue for me. I guess I'll have to just use top to get CPU utilization stats, because the numbers from uptime are not directly related to that. 🙂
 
Originally posted by: Nothinman
However, when you run intense applications such as SQL server, .NET applictions server, etc... and do some sort of productive work on your computer, it is normal for computer to misbehave once in a while

I can understand your workstation having problems, because you're fiddling with things all the time, but unless you're breaking your SQL server frequently it should run flawlessly no matter how much CPU time it consumes constantly. The fact that it's doing a lot of calculations and/or I/O does not make it more prone to breakage, well atleast not with decent software =)

to end the string of long quotes because it's annoying as all hell...


I agree, it's understandable if you reboot your workstation periodically however the server really should have very little downtime whatsoever.

It all comes down to the bane of windows operating systems: memory leaks. I beleive that is the major reason that the average windows server or desktop is not able to keep up with Unices and Netware OS.
No, No, No, No, No; if you were comparing Windows 98 than I would agree 100%, but if you are running NT/2K/XP the only "memory leaks" you should get is from a bad application, even than you can just end the task and since there is seperate memory space you get your RAM back. If there is a service that has a memory leak you can restart the service, etc. however your comparison is just not valid. If you watch the amount of RAM used in taskmgr it will increase some over time since the OS is casching some of the dlls used in RAM (to make thigns go quicker next time you need them) however this is not a memory leak.

-Spy
 
I have to say, in my experience Windows isn't too good at handling long term heavy workstation use.

My box at work(I have a Windows box so I can use Outlook, FW-1 mgmt tools, and some other stuff) runs fine for a while, but I do have to reboot it once in a while(we're talking weeks of uptime though, so no disaster).

This ain't no biggie, it's stable enough for me, but still, I should never HAVE to reboot my computer unless I feel like it for some reason.
 
It's not that Unices have such a superior memory control. I was just remarking on the kneejerk reaction of windows admins and why that leads to much less uptime then the average unix or netware box.

Memory leaks are a problem with any OS. The solution for a quick fix for in the windows world is just to reboot(almost trained to from birth with win95 and win98 experiance to grow off of), because if there are anything wrong there is not a whole lot you can do to solve the problem. As w2k/xp becomes more Unix-like (ie mutiple users, "task manager" *cough* TOP ripoff..., and more effective remote administration ability etc etc) the differences between the 2 OS's become more blurred, but as a user/admin I just appreciate a OS that is flexible and I have a wide selection of solutions that I can talor for specific situtations, thats why I prefer Linux. Also why would I want to spend 600+ dollars for a crippled version of window server (or about 1500 for advanced server and still be limited to 10 cleint computers) if I can just download gentoo or debain for free and donate some money to a good cause?
 
Originally posted by: Sunner
I have to say, in my experience Windows isn't too good at handling long term heavy workstation use.

My box at work(I have a Windows box so I can use Outlook, FW-1 mgmt tools, and some other stuff) runs fine for a while, but I do have to reboot it once in a while(we're talking weeks of uptime though, so no disaster).

This ain't no biggie, it's stable enough for me, but still, I should never HAVE to reboot my computer unless I feel like it for some reason.
How do you know that's a fault of windows and not a 3rd party driver and/or software?

Dont forget the real reason that we all restart for, hardware changes 😎

-Spy
 
BTW just when it comes to us "amateurs" dogging on window's security ability. Just don't forget that it took 6 years of active use for hackers find a single known exploitable flaw in OpenBSD's stable releases and XP was hacked even before it was ever released, but then again we all know that closed source developement is inherintly more secure than opensource.
 
Originally posted by: spyordie007
Originally posted by: Sunner
I have to say, in my experience Windows isn't too good at handling long term heavy workstation use.

My box at work(I have a Windows box so I can use Outlook, FW-1 mgmt tools, and some other stuff) runs fine for a while, but I do have to reboot it once in a while(we're talking weeks of uptime though, so no disaster).

This ain't no biggie, it's stable enough for me, but still, I should never HAVE to reboot my computer unless I feel like it for some reason.
How do you know that's a fault of windows and not a 3rd party driver and/or software?

Dont forget the real reason that we all restart for, hardware changes 😎

-Spy

I thought it was crappy rural power. 😕
 
Originally posted by: drag
BTW just when it comes to us "amateurs" dogging on window's security ability. Just don't forget that it took 6 years of active use for hackers find a single known exploitable flaw in OpenBSD's stable releases and XP was hacked even before it was ever released, but then again we all know that closed source developement is inherintly more secure than opensource.

ISS employees dont count as hackers.
 
Back
Top