I'm now trying http_load and it seems to spread the load quite equally
according to ipvsadm output. So what is the deal with Netscrap? Can
someone explain it to me.
-jeremy
>
> I finally got an environment available to me to really setup Ultramonkey
> and let me tell you, it's sweet! Works like a champ. I only noticed one
> thing that seemed a bit disturbing.
>
> On the real servers I'm accumulating quite a bit of TIME_WAIT connections
> which I assume is from ldirectord checking the status page on the web
> server. Is this a problem? Should I change my checkinterval?
>
> Also another thing is that I'm using round robin and lvs isn't acting very
> rr like. Basically I'm just hitting shift reload on my browser and
> expecting it to say web1, web2, web3 when hitting the different servers
> but it doesn't always happen. Problem is I can't tell if it's my browser
> or LVS. If I go to an invalid link and then hit the back button, it ends
> of on another server. But at one point I did have it nicely rotating on
> shift reloads, so I'm a bit confused on this.
>
> Other then that, the failover and monitoring of real servers has been
> perfect thus far. I'm very happy!
>
> -jeremy
>
>
>
--
http://www.xxedgexx.com | jeremy@xxxxxxxxxxxx
---------------------------------------------
|