Performance

[webman]

Pretty graph

 

The first spike is immediately after they start the first test. Has anyone else monitored the status of the server this way while running them and what sort of usage do you get?

 

Our server is a P4 1.8GHz, 1GB RAM, EIDE hard disk and a 1Gbps NIC.

[Disease]

No not had anything like that yet, but we do run the APS on a twin xeon 3.0 with 3 gig of Ram

[sahmeepee]

You have to be very careful with the default graph scales in perfmon. In your graph for example, the top of the graph isn't representing 1gbps or even 100mbps. The graphs DO NOT pick a sensible scale automatically for many counter types!

 

If you look at the stats in text just under the graph, your peak bytes received/sec is only 3,779,984 (= 30,239,872bps, =~30Mbps) Even if your switches are running at 100Mbps, you're not pushing your full bandwidth.

[webman]

Yeah, I know it's in percentages, so doesn't give the best view. I remember finding a good netstat graph app ages ago but can't remember the name of it :? It seems to be coping OK though, 90 users at once today.

[ChrisH]

I have a similar spec machine to you webman and monitored it carefully today with process explorer which I was introduced to yeasterday by some french bloke I noticed the spiking only occured for the processor as they where all logging on and during and at the end of the test everything remained stable. There wasnt much utilisation on the network front from what I could see.