Alphatran wrote:
Andy Colson wrote:
First, I think we need to find the bottleneck.
a) If there is only one client, requesting one map, then you will not
have enough data going over the line to use 200Mbps.
b) If you have lots of clients, all requesting lots of maps, then you'll
have enough data going over the line to use 200Mbps.
Assuming b is true, then we need to see if you are cpu bound or io
bound. On you linux box watch "vmstat 5" for a while (you could even
post the results if you want).
The last column "wa" is %cpu wait time for the hard drve. If that's
high then you're io bound.
Next column in is "id", or %cpu idle time. If that's near zero then you
are cpu bound.
-Andy
Thanks Andy,
I have just tested with 2 clients. There is one map with many layers. These
lines are result of "vmstat 5" command, please help me!
[r...@gis ~]# vmstat 5
procs -----------memory---------- ---swap-- -----io---- --system--
-----cpu------
r b swpd free buff cache si so bi bo in cs us sy id
wa st
0 0 31368 14268 11100 1450188 1 2 428 46 53 66 2 0 96
2 0
0 0 31368 14396 11108 1450188 0 0 0 17 184 197 0 0 100
0 0
0 0 31368 14388 11116 1450188 0 0 0 22 167 197 0 0 100
0 0
0 0 31368 14468 11124 1450188 0 0 0 12 182 202 0 0 100
0 0
0 0 31368 14468 11124 1450188 0 0 0 20 165 196 0 0 100
0 0
Its a little hard to read wrapped like that, but, looks to me like cpu
is 100% idle. Looks like you have an extra "st" column, I'm not sure
what that is.
With two clients, how was the response time? You are not going to
generate very much data, so you wont use very much network bandwidth.
200Mbps / 8 bits = 25 Meg per second, a mapfile is going to generate
10-20k, which is what, 1% of your bandwidth. If response time on the
client was ok, then I'm not sure what we are trying to solve.
-Andy
_______________________________________________
mapserver-users mailing list
mapserver-users@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/mapserver-users
