Guys,
I need some input to a server problem our leagues have been
experiencing. At some point during the race, all cars disappear, and
the latency meter goes full red. This lasts about 15 seconds, and then
everything is fine for the rest of the event. The most common time
this happens is right as all cars leave pit road at the start of the
event.
I initially thought this was a network issue. But it's happened at 2
different host locations (ISP's). Since this usually happens right at
the start of the race, when the server is at it's busiest, could this
be a problem with the machine itself?
Yesterday I ran a test event, and ran a system monitor in the
background. I checked for things like processor usage, cache errors,
swap file usage, and physical memory usage. Everything looked good in
the reports, except that the processor ran at 99-100% constant from
the time drivers started joining the session.
I'm curious if any other server owners have seen this, and if you have
any indication of the problem. Our server is configured as follows:
Celeron 900
Asus CUSL2-C motherboard
3COM 3CR990 Nic
384mb ram
500mb fixed swap file
Win98 SE
Tim Wortman
North American Simulation Series
www.nasscar.com
Celery 900 is most likely the problem.....But you didn't mention hard drive space, or what type of
Internet connection you were using.
--
Biz
"Don't touch that please, your primitive intellect wouldn't understand
alloys and compositions and,......things with molecular structures,....and
the....." - Ash
TimW <[email hidden]> wrote in message news:[email hidden]...
> Guys,
>
> I need some input to a server problem our leagues have been
> experiencing. At some point during the race, all cars disappear, and
> the latency meter goes full red. This lasts about 15 seconds, and then
> everything is fine for the rest of the event. The most common time
> this happens is right as all cars leave pit road at the start of the
> event.
>
> I initially thought this was a network issue. But it's happened at 2
> different host locations (ISP's). Since this usually happens right at
> the start of the race, when the server is at it's busiest, could this
> be a problem with the machine itself?
>
> Yesterday I ran a test event, and ran a system monitor in the
> background. I checked for things like processor usage, cache errors,
> swap file usage, and physical memory usage. Everything looked good in
> the reports, except that the processor ran at 99-100% constant from
> the time drivers started joining the session.
>
> I'm curious if any other server owners have seen this, and if you have
> any indication of the problem. Our server is configured as follows:
>
> Celeron 900
> Asus CUSL2-C motherboard
> 3COM 3CR990 Nic
> 384mb ram
> 500mb fixed swap file
> Win98 SE
>
>
>
> Tim Wortman
> North American Simulation Series
> www.nasscar.com
It's got a 10gig 5400rpm drive with roughly 75% of the drive free. Box
is on a T3 line at an ISP.
On Sun, 05 May 2002 17:01:09 GMT, "Biz" <[email hidden]>
wrote:
Tim Wortman
North American Simulation Series
www.nasscar.com
Quoted text (48 lines)
>Celery 900 is most likely the problem.....But you didn't mention hard drive space, or what type of >Internet connection you were using. >-- >Biz > >"Don't touch that please, your primitive intellect wouldn't understand >alloys and compositions and,......things with molecular structures,....and >the....." - Ash > > >TimW <[email hidden]> wrote in message news:[email hidden]... >> Guys, >> >> I need some input to a server problem our leagues have been >> experiencing. At some point during the race, all cars disappear, and >> the latency meter goes full red. This lasts about 15 seconds, and then >> everything is fine for the rest of the event. The most common time >> this happens is right as all cars leave pit road at the start of the >> event. >> >> I initially thought this was a network issue. But it's happened at 2 >> different host locations (ISP's). Since this usually happens right at >> the start of the race, when the server is at it's busiest, could this >> be a problem with the machine itself? >> >> Yesterday I ran a test event, and ran a system monitor in the >> background. I checked for things like processor usage, cache errors, >> swap file usage, and physical memory usage. Everything looked good in >> the reports, except that the processor ran at 99-100% constant from >> the time drivers started joining the session. >> >> I'm curious if any other server owners have seen this, and if you have >> any indication of the problem. Our server is configured as follows: >> >> Celeron 900 >> Asus CUSL2-C motherboard >> 3COM 3CR990 Nic >> 384mb ram >> 500mb fixed swap file >> Win98 SE >> >> >> >> Tim Wortman >> North American Simulation Series >> www.nasscar.com > >Can you check the tracking of drivers accessing the system when the lag
happens? It could be a client or clients on a slower machine(s) bogging
down the server at the point of gaining access. Any common denominator at
that point, since it happens on a different ISP?
--
Chris H.
"TimW" <[email hidden]> wrote in message
news:[email hidden]...
> It's got a 10gig 5400rpm drive with roughly 75% of the drive free. Box
> is on a T3 line at an ISP.
>
> On Sun, 05 May 2002 17:01:09 GMT, "Biz" <[email hidden]>
> wrote:
>
> >Celery 900 is most likely the problem.....But you didn't mention hard
drive space, or what type of
> >Internet connection you were using.
> >--
> >Biz
> >
> >"Don't touch that please, your primitive intellect wouldn't understand
> >alloys and compositions and,......things with molecular
structures,....and
> >the....." - Ash
> >
> >
> >TimW <[email hidden]> wrote in message
news:[email hidden]...
> >> Guys,
> >>
> >> I need some input to a server problem our leagues have been
> >> experiencing. At some point during the race, all cars disappear, and
> >> the latency meter goes full red. This lasts about 15 seconds, and then
> >> everything is fine for the rest of the event. The most common time
> >> this happens is right as all cars leave pit road at the start of the
> >> event.
> >>
> >> I initially thought this was a network issue. But it's happened at 2
> >> different host locations (ISP's). Since this usually happens right at
> >> the start of the race, when the server is at it's busiest, could this
> >> be a problem with the machine itself?
> >>
> >> Yesterday I ran a test event, and ran a system monitor in the
> >> background. I checked for things like processor usage, cache errors,
> >> swap file usage, and physical memory usage. Everything looked good in
> >> the reports, except that the processor ran at 99-100% constant from
> >> the time drivers started joining the session.
> >>
> >> I'm curious if any other server owners have seen this, and if you have
> >> any indication of the problem. Our server is configured as follows:
> >>
> >> Celeron 900
> >> Asus CUSL2-C motherboard
> >> 3COM 3CR990 Nic
> >> 384mb ram
> >> 500mb fixed swap file
> >> Win98 SE
> >>
> >>
> >>
> >> Tim Wortman
> >> North American Simulation Series
> >> www.nasscar.com
> >
> >
>
>
>
> Tim Wortman
> North American Simulation Series
> www.nasscar.com
>
On Sun, 05 May 2002 18:09:01 GMT, "Chris H." <[email hidden]>
wrote:
>Can you check the tracking of drivers accessing the system when the lag
>happens? It could be a client or clients on a slower machine(s) bogging
>down the server at the point of gaining access. Any common denominator at
>that point, since it happens on a different ISP?
I guess I could somehow. But what concerned me was the 100% processor
occupancy rate. That tells me maybe the processor can't keep up with
what's being asked of it.
I thought the processor load for a server machine wasn't that severe.
Guess I was wrong.
Tim Wortman
North American Simulation Series
www.nasscar.com
See the N2002 RAS Open Server thread below, but I am running the following:
Intel VC820 i820 chipset slot one board
Pentium 3 550 non coppermine 512k cache
2 x 128MB PC600 Rambus RDRAM
18% CPU time is the max server load with 15 clients that I have seen in
dedicated server mode, direct3d renderer 800x600x16 bit running in a window
with screen res at 1024x768x16bit
This is Windows XP Pro run in windows classic mode with all of the
enhancements turned off, all unnecessary services stopped, no reomte
desktop, no system restore. Every last thing I can find is off.
It may help to bring up task manager and make the server window not the
active window with 'foucs'. Leave the focus on the task manager, and I see
CPU load with 15 clients at 4-6%.
jason
jason -A-T- jason fagan -D-O-T- com
Nascar 2002 Server
Name: ras open server
Access Password: -none-
Boss Password: ras
See the N2002 RAS Open Server thread below, but I am running the following:
Intel VC820 i820 chipset slot one board
Pentium 3 550 non coppermine 512k cache
2 x 128MB PC600 Rambus RDRAM
18% CPU time is the max server load with 15 clients that I have seen in
dedicated server mode, direct3d renderer 800x600x16 bit running in a window
with screen res at 1024x768x16bit
This is Windows XP Pro run in windows classic mode with all of the
enhancements turned off, all unnecessary services stopped, no reomte
desktop, no system restore. Every last thing I can find is off.
It may help to bring up task manager and make the server window not the
active window with 'foucs'. Leave the focus on the task manager, and I see
CPU load with 15 clients at 4-6%.
jason
jason -A-T- jason fagan -D-O-T- com
Nascar 2002 Server
Name: ras open server
Access Password: -none-
Boss Password: ras
In article <[email hidden]>,
[email hidden] says...
By default, the server runs with high priority. If you want to reduce
CPU load, pass in the following parameter on the command line:
-kh0
That's a zero at the end.
This will allow the server to run in normal priority mode.
--
=========================================================
DeMONS/1 for Nascar Racing 3 & Nascar Legends
http://www.paddedwall.org/demons
DeMONS/2 for Nascar Racing 4 and 2002 Season (in development)
http://www.paddedwall.org/demons2
RASCAR Roster
http://www.paddedwall.org/rascar
Redneck Techno-Biker & "programming deity"
http://www.paddedwall.org/john
Barbarian Diecast Collector (460+ cars and counting)
http://www.paddedwall.org/diecast
If you want to send me email, go to the first URL shown
above & click "Send Me Mail" in the contents frame.
=========================================================
Quoted text (14 lines)
> On Sun, 05 May 2002 18:09:01 GMT, "Chris H." <[email hidden]> > wrote: > > >Can you check the tracking of drivers accessing the system when the lag > >happens? It could be a client or clients on a slower machine(s) bogging > >down the server at the point of gaining access. Any common denominator at > >that point, since it happens on a different ISP? > > I guess I could somehow. But what concerned me was the 100% processor > occupancy rate. That tells me maybe the processor can't keep up with > what's being asked of it. > > I thought the processor load for a server machine wasn't that severe. > Guess I was wrong.Like I mentioned above, its problably the Celery 900. Most dedicated servers run under 25-40% cpu
usage. And remember, for the most part a Celeron is a Pentium 1.5 if you will. They weren't as
good as most PII's, but better than Pentiums.
--
Biz
"Don't touch that please, your primitive intellect wouldn't understand
alloys and compositions and,......things with molecular structures,....and
the....." - Ash
TimW <[email hidden]> wrote in message news:[email hidden]...
> On Sun, 05 May 2002 18:09:01 GMT, "Chris H." <[email hidden]>
> wrote:
>
> >Can you check the tracking of drivers accessing the system when the lag
> >happens? It could be a client or clients on a slower machine(s) bogging
> >down the server at the point of gaining access. Any common denominator at
> >that point, since it happens on a different ISP?
>
> I guess I could somehow. But what concerned me was the 100% processor
> occupancy rate. That tells me maybe the processor can't keep up with
> what's being asked of it.
>
> I thought the processor load for a server machine wasn't that severe.
> Guess I was wrong.
>
>
> Tim Wortman
> North American Simulation Series
> www.nasscar.com
Sounds like something is wrong. :(
My league server is a P3-800, and with ***three*** N2002 hosts running
simultaneously, the processor is at 38-45% usage.
Are you using the -kh0 switch in your command line?
What's your processor usage without N2002 running?
-Paul Eigsti
-League & Server admin
Western States Racing League
www.wsrlracing.com
My e-mail address is false.
Please reply to this news group.
Without N2002 server running, processor usage is 3-8%.
On Sun, 5 May 2002 16:41:27 -0700, "Paul E" <[email hidden]> wrote:
Tim Wortman
North American Simulation Series
www.nasscar.com
Quoted text (18 lines)
>Sounds like something is wrong. :( > >My league server is a P3-800, and with ***three*** N2002 hosts running >simultaneously, the processor is at 38-45% usage. > >Are you using the -kh0 switch in your command line? > >What's your processor usage without N2002 running? > >-Paul Eigsti >-League & Server admin >Western States Racing League >www.wsrlracing.com >My e-mail address is false. >Please reply to this news group. > > >On Sun, 05 May 2002 19:52:35 GMT, John Simmons <[email hidden]> wrote:
>By default, the server runs with high priority. If you want to reduce
>CPU load, pass in the following parameter on the command line:
>
> -kh0
No where in the readme is this documented. I've always ran with the
command line of -k0, but not -kh0.
Tim Wortman
North American Simulation Series
www.nasscar.com
The "-kh0" command line option is specifically mentioned in the readme in
reference to running the game at lower priority if it's interfering with
chat programs such as Microsoft's GameVoice. I'm not sure where you got it
from, but AFAIK "-k0" is not a valid command line option and should have no
affect on the game.
- Eric
"TimW" <[email hidden]> wrote in message
news:[email hidden]...
> No where in the readme is this documented. I've always ran with the
> command line of -k0, but not -kh0.
Are you running the server minimized? I have had problems like this before
and then switched the windows from being minimized to not being minimized
and fixed the problem.
Bill / Amish on Team Brickyard
www.iracingfx.com
"TimW" <[email hidden]> wrote in message
news:[email hidden]...
Quoted text (35 lines)
> Guys, > > I need some input to a server problem our leagues have been > experiencing. At some point during the race, all cars disappear, and > the latency meter goes full red. This lasts about 15 seconds, and then > everything is fine for the rest of the event. The most common time > this happens is right as all cars leave pit road at the start of the > event. > > I initially thought this was a network issue. But it's happened at 2 > different host locations (ISP's). Since this usually happens right at > the start of the race, when the server is at it's busiest, could this > be a problem with the machine itself? > > Yesterday I ran a test event, and ran a system monitor in the > background. I checked for things like processor usage, cache errors, > swap file usage, and physical memory usage. Everything looked good in > the reports, except that the processor ran at 99-100% constant from > the time drivers started joining the session. > > I'm curious if any other server owners have seen this, and if you have > any indication of the problem. Our server is configured as follows: > > Celeron 900 > Asus CUSL2-C motherboard > 3COM 3CR990 Nic > 384mb ram > 500mb fixed swap file > Win98 SE > > > > Tim Wortman > North American Simulation Series > www.nasscar.comCelery wouldn't make a difference.
Bill / Amish on Team Brickyard
www.iracingfx.com
"Biz" <[email hidden]> wrote in message
news:pzdB8.7433$[email hidden]...
> Celery 900 is most likely the problem.....But you didn't mention harddrive space, or what type of
news:[email hidden]...
Quoted text (10 lines)
> Internet connection you were using. > -- > Biz > > "Don't touch that please, your primitive intellect wouldn't understand > alloys and compositions and,......things with molecular structures,....and > the....." - Ash > > > TimW <[email hidden]> wrote in messageQuoted text (37 lines)
> > Guys, > > > > I need some input to a server problem our leagues have been > > experiencing. At some point during the race, all cars disappear, and > > the latency meter goes full red. This lasts about 15 seconds, and then > > everything is fine for the rest of the event. The most common time > > this happens is right as all cars leave pit road at the start of the > > event. > > > > I initially thought this was a network issue. But it's happened at 2 > > different host locations (ISP's). Since this usually happens right at > > the start of the race, when the server is at it's busiest, could this > > be a problem with the machine itself? > > > > Yesterday I ran a test event, and ran a system monitor in the > > background. I checked for things like processor usage, cache errors, > > swap file usage, and physical memory usage. Everything looked good in > > the reports, except that the processor ran at 99-100% constant from > > the time drivers started joining the session. > > > > I'm curious if any other server owners have seen this, and if you have > > any indication of the problem. Our server is configured as follows: > > > > Celeron 900 > > Asus CUSL2-C motherboard > > 3COM 3CR990 Nic > > 384mb ram > > 500mb fixed swap file > > Win98 SE > > > > > > > > Tim Wortman > > North American Simulation Series > > www.nasscar.com > >N4/2002 is memory intensive vs. cpu intensive.
Bill / Amish on Team Brickyard
www.iracingfx.com
"TimW" <[email hidden]> wrote in message
news:[email hidden]...
> On Sun, 05 May 2002 18:09:01 GMT, "Chris H." <[email hidden]>
> wrote:
>
> >Can you check the tracking of drivers accessing the system when the lag
> >happens? It could be a client or clients on a slower machine(s) bogging
> >down the server at the point of gaining access. Any common denominatorat
Quoted text (13 lines)
> >that point, since it happens on a different ISP? > > I guess I could somehow. But what concerned me was the 100% processor > occupancy rate. That tells me maybe the processor can't keep up with > what's being asked of it. > > I thought the processor load for a server machine wasn't that severe. > Guess I was wrong. > > > Tim Wortman > North American Simulation Series > www.nasscar.comAssuming the "normal priority switch" doesn't help...
Something in the TCP/IP part of the OS causing the CPU to max out? Flaky NIC?
I, too, think the Celeron should be plenty of CPU.
Please post if you figure out what's going on.
Good luck,
Gerald
[email hidden] (TimW) wrote in message news:<[email hidden]>...
