|
Hi
On Tue 07 Feb 2006 16:54:17 GMT , Randy Paries <rtparies@xxxxxxxxx> wrote:
Does this give you any insight?
ipvsadm -Ln --rate
IP Virtual Server version 1.0.10 (size=65536)
Prot LocalAddress:Port CPS InPPS OutPPS InBPS OutBPS
-> RemoteAddress:Port
TCP 66.208.187.10:80 61 1220 0 108590 0
-> 192.168.0.159:80 52 867 0 72101 0
-> 192.168.0.204:80 5 138 0 16551 0
-> 192.168.0.203:80 4 214 0 19937 0
TCP 66.208.187.10:443 0 0 0 0 0
-> 192.168.0.203:443 0 0 0 0 0
192.168.0.159 - I assume this is local.lovejoy? If so, it's handling 10
times the connections per second that the other are for a given number
of ActiveConn in the basic output.
ipvsadm -Ln --stats
IP Virtual Server version 1.0.10 (size=65536)
Prot LocalAddress:Port Conns InPkts OutPkts InBytes OutBytes
-> RemoteAddress:Port
TCP 66.208.187.10:80 22291846 1257M 0 152G 0
-> 192.168.0.159:80 742593 8850902 0 998205K 0
-> 192.168.0.204:80 10094273 593304K 0 71634M 0
-> 192.168.0.203:80 11422135 655111K 0 80084M 0
TCP 66.208.187.10:443 117751 993403 0 155060K 0
-> 192.168.0.203:443 49549 419180 0 66259192 0
You added local.lovejoy after the other machines by running ipvsadm by
hand, right?
The "--rate" output makes it look to me that your new server is a *lot*
better in terms of performance when compared to your other two machines.
As I remember it, the wlc scheduler looks at the number of active
connections and sends the next connection to the least-loaded server.
Given that your ipvsadm output is only a specific point in time, and it
shows that local.lovejoy has *more* connections than the other two,
that might not seem too sensible - however if those connections are
getting to TIME_WAIT or FIN_WAIT much more quickly than the other two
machines, it is able to handle more connections and therefore the
real-time figure is likely to bounce up and down around the same level
as the older machines.
What specifications are your older machines, and what is the new one?
What Apache versions are they running?
Graeme
|
