Speed collapses when Traffic is rising

[blowspark]

Hi,

The Specs:

P4 3ghz HT - FSB 800MHz
1GB ECC DDR-SDRAM @ 400MHz
3Ware AMCC 9500S - SATA Raid Card
4 x 250GB Seagate Barracuda 7200.8 SATA NCQ

OS: Windows 2k3 - SP2


ioFTPD version : 5-8-5
ioBanana version: 20


When I test the Raid Speed without any Traffic I get an average Read Speed = 74 MB /s.


See this Pic here :


A speed that satisfies, with crypt, i think.

But now the Problem:

The Speed collapses the more Users are connected and cause Traffic.

On the next Picture you will see what I mean :


At this time there are only 4 Users connected causing a Traffic of ~ 10 Mbit/s and as you can see the average Speed

decreases horrible. Only 40.8 MB/s average Read Speed, and the Graph fluctuate horrible.

And this resumes as more Users are connected.

Here is a Screen what it looks like when ~ 20 Users are connected causing a Traffic of ~ 25 Mbit/s :


Average Speed sinks down to 6.5 MB /s and the System begins to lag,

but the maximum CPU Usage is between 20 and 40 % , that should ´nt be the Problem.




Maybe someone has a suggestion for this problem.

[_panic_]

wow, thanks for such a thorough post.

let me see if i understand, since there is a lot going on here.

you're running a benchmark tool called hd tac. when you run this benchmark with no users, you get the first graph (~74MB/sec).

in your second graph, with ten users connected, you run hd tac again, and generate the middle benchmark statistics (~40MB/sec). the users themselves are using ~10MB/sec, so we can't account for ~24MB/sec of capacity somewhere.

when the users begin to generate ~25MB/sec of usage, your benchmark drops to 6.5MB/sec, so we have lost ~40MB/sec of capacity somewhere.

does that fairly resummarize your problem?

i have a couple questions. are the users some kind of "control", or are they just random users connected? i ask because i'm trying to figure out how much of this issue is a controlled measurement vs how much is random chance. this is important for i/o contention. (the pattern should look different if 10 people are grabing the same file, or if they all are getting different files.)

second, are you experiencing a problem that caused you to want to run a benchmark, or are did you get concerned after you saw the benchmark result? what does running two hd tac programs at the same time report? i'm trying to figure whether the issue is ioftpd, or whether ioftpd is a canary for a deeper configuration/contention problem.

[ADDiCT]

Is this one big raid-0 array, or just disk spanning?