|
On Fri, 3 Mar 2000, Michael Sparks wrote:
>
> We're doing stuff like this on our servers at the moment - using a bunch
> of perl programs I cobbled together - the real servers themselves detect
> system overload from a variety of system atrtributes, and based on that
> tell a program on the LVS load balancer whether they're overloaded/
> underloaded/ nominal, which then adjusts the weights of teh real servers
> accordingly.
don't reinvent the wheel. There's the bproc file system, or proc for
beowulfs, which make proc for all machines in a cluster visible
to the other machines
Joe
>
>
>
> Michael
> --
> National & Local Web Cache Support R: G117
> Manchester Computing T: 0161 275 7195
> University of Manchester F: 0161 275 6040
> Manchester UK M13 9PL M: Michael.Sparks@xxxxxxxxxxxxxxx
>
>
>
>
--
Joseph Mack mack@xxxxxxxxxxx
|
