|
Hi,
On Thu, Aug 21, 2008 at 05:57:31PM -0400,
Techiie <techiie@xxxxxxxxxxx> wrote:
> I cannot seem to get the configuration nailed down for the https side. http
> works fine. If I don't use a cron job to reload it, then the only https
> requests that work are the ones that go to 192.168.5.13. the rest have a
> weight of zero. I have tried changing it from gated to wlc to rr and
> nothing seems to work. Any ideas are greatly appreciated.
>
> Prot LocalAddress:Port Scheduler Flags
> -> RemoteAddress:Port Forward Weight ActiveConn InActConn
> TCP www.webpage:http wlc
> -> server1:htt Route 1 0 70
> -> server2:htt Local 1 1 56
> -> server3:htt Route 1 1 60
> TCP www.webapge:https wlc
> -> server1:htt Route 1 0 0
> -> server2:htt Local 1 0 0
> -> server3:htt Route 0 0 0
ldirectord seems to set your real servers "quiescent" == weight 0.
I assume health checking simply fails for these real servers. Check,
manually from the director, whether you can fetch the specified URL
(html-document) and verify the result you/ldirectord (receive parameter)
expects. You should find some notifications in the log of ldirectord,
as well.
> The above only runs because I used a cron job to reload the ipvsadm as
> follows:
>
> -A -t www.webapge.com:http -s wlc
> -a -t www.webapge.com:http -r server1:http -g -w 1
> -a -t www.webapge.com:http -r server2:http -g -w 1
> -a -t www.webapge.com:http -r server3:http -g -w 1
> -A -t www.webapge.com:https -s wlc
> -a -t www.webapge.com:https -r server1:https -g -w 0
> -a -t www.webapge.com:https -r server2:https -g -w 1
> -a -t www.webapge.com:https -r server3:https -g -w 1
This would help, temporarily, but does not prevent ldirectord from taking
them out of ipvs during the next health checking cycle.
> # HTTPS Virtual Service
> virtual=192.168.5.6:443
> real=192.168.5.11:443 gate 1
> real=192.168.5.12:443 gate 1
> real=192.168.5.13:443 gate 1
> service=https
> request="lvscheck.html"
> receive="lvscheck"
> virtualhost="www.webpage.com"
> scheduler=wlc
> protocol=tcp
Regards,
Volker
|
