Hello,
On Wed, 31 Oct 2001, Nicolas Chiappero wrote:
> Hello,
>
> Can you explain the possible reasons that can
> cause an ipvsadm ouput looking like this :
>
> IP Virtual Server version 0.9.7 (size=4096)
> Prot LocalAddress:Port Scheduler Flags
> -> RemoteAddress:Port Forward Weight ActiveConn InActConn
> TCP XXX.XXX.XXX.XXX:443 rr
> -> 172.22.1.105:443 Route 1 7 41
> -> 172.22.1.107:443 Route 1 3 44
> -> 172.22.1.104:443 Route 1 9 44
> -> 172.22.1.103:443 Route 1 8 38
> TCP YYY.YYY.YYY.YYY:80 rr
> -> 172.22.1.107:80 Route 1 25 1008
> -> 172.22.1.103:80 Route 1 28 1000
> -> 172.22.1.104:80 Route 1 25 1008
> -> 172.22.1.105:80 Route 1 36 993
> TCP XXX.XXX.XXX.XXX:80 rr
> -> 172.22.1.103:80 Route 1 303 10573
> ^^^
You have to check the web server log. May be your kernel
or the web server is not configured for many processes or file
descriptors? The scheduler has performed its role, the requests
were sent to the real server but the real server does not serve
the connections as expected.
> -> 172.22.1.105:80 Route 1 1392 9508
> -> 172.22.1.104:80 Route 1 1273 9188
> -> 172.22.1.107:80 Route 1 1330 9572
>
> This director is a Linux box running a 2.2.14 kernel.
> I cannot figure why 172.22.1.103 realserver is not having
> the same number of active connections although I use round
> robin load balancing and same weight on each realserver.
>
> Nicolas.
Regards
--
Julian Anastasov <ja@xxxxxx>
|