[f-nsp] Response-time balancing

[Vlad]

thanks for your reply. I can't have SI to mark real server as failed,
cause we need sticky sessions on SSL port same time. But documentation
[http://www.foundrynetworks.co.jp/services/documentation/siug/ServerIron_Server_Load_Balancing.html]
says:

<quote>
NOTE: For SwitchBack (Direct Server Return) configurations, since the
ServerIron does not see the server reply traffic, the ServerIron uses
only the health check responses to measure the response time.
</quote>

so this is what my assumption was based on when I created a custom
health check script that delays its response according to real server
load....

-
Vlad




On Tue, Nov 25, 2008 at 3:13 AM, Mischa Peters <foundry at high5.nl> wrote:
> Hi Vlad,
>
> The weight that you can add to a real server only applies to the predictor
> weighted.
> When you apply the predictor least connections, you can not configure
> anything at real server level.
> The url that you are using only applies to health checks. If you want to tie
> this is with the load balancing mechanism you need to create an http match
> list.
>
> Something like:
>
> http match-list hc
>  default down
>  up simple UP
>  down simple DOWN
>
> If you then tie this together at the real server like:
>
> port http keepalive
> port http url "GET /cgi-bin/la.cgi"
> port http content-match hc
>
> If you have your script present a DOWN when it's over it's limit and an UP
> when it's not you will get what you want from the script.
> When the script presents DOWN it leave existing connections alone but this
> server will not be used for the next incoming connection.
>
> Hope this helps.
>
> Mischa
>