[K12OSN] Re: tuxpaint and others running slow
Dan Kegel
daniel.r.kegel at gmail.com
Sat Oct 22 05:21:47 UTC 2005
On 10/21/05, Justin Paulsen <paulsenj at frederic.k12.wi.us> wrote:
> I am having a similar problem, as current I've only experienced this
> with TuxType2 and VMware (not a surprise there). I'm fine with a few
> users but at about 14 or so its unusable.
>
> Server: 1x AMD Opteron 252 1.8GHz (Dual proc capable)
> 2x 160GB 7200 RPM SATA drives in RAID 1
> Dual Gbit NIC
> 4GB RAM
>
> Switch: 24pt Gbit (D-Link)
Let's figure out what the bottleneck is. (This goes for anyone who's having
performance problems on an LTSP server.)
Run just one of the two apps, say TuxType2, on all workstations at once.
Verify that the slowdown happens.
Once you have the slowdown going, then let's look at the server:
run 'top' on the server. Grab the output with the mouse
and post it here so we can see what's up. It'll look something like this:
top - 22:24:23 up 49 min, 2 users, load average: 0.10, 0.09, 0.07
Tasks: 86 total, 1 running, 84 sleeping, 0 stopped, 1 zombie
Cpu(s): 2.6% us, 0.3% sy, 0.0% ni, 96.7% id, 0.0% wa, 0.3% hi, 0.0% si
Mem: 515304k total, 309264k used, 206040k free, 15984k buffers
Swap: 1048568k total, 0k used, 1048568k free, 178104k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
3221 root 15 0 100m 17m 88m S 2.0 3.4 0:49.79 X
3572 dank 15 0 37536 12m 21m S 0.3 2.4 0:03.33 gnome-terminal
...
That'll tell us a lot.
And let's look at the network:
use ping to examine the health of the network connecting the server to
the workstations, e.g. run "ping -c100 -A client1" on the server,
and post the statistics it prints at the end. It'll look like this:
100 packets transmitted, 100 received, 0% packet loss, time 19899ms
rtt min/avg/max/mdev = 0.493/0.512/0.553/0.015 ms, pipe 2, ipg/ewma
201.002/0.513 ms
Any packet loss indicates trouble with the network.
I'm looking forward to seeing your results.
More information about the K12OSN
mailing list