|
the differences beetwen the two request (VIP or RIP) should be only the
target address, maybe you have differente configuration in your
apache.conf to serve requested to RIP and VIP (maybe a virtual host with
logging disabled or pointing to a dedicated access_log).
Agostino
--
Experience is when you get what you don't want to get!
On Thu, 6 Nov 2008 at 13:33, Julien Cornuwel wrote:
> All three servers are Ubuntu 8.04.
>
> The load balancer is configured to use lc algo and direct routing on the web
> servers.
> Network :
> eth0 - RIP1
> eth0:1 - VIP
>
> Each webserver (apache2) has the following parameters set in
> /etc/sysctl.conf :
> net.ipv4.conf.eth0.arp_ignore = 1
> net.ipv4.conf.eth0.arp_announce = 2
> Network :
> eth0 - RIP2 & RIP3
> lo:0 - VIP
>
> The cluster has been working fine for 3 weeks now. But now I'm willing to do
> some stats and here is what I have in my access_log :
> 127.0.0.1 - - [06/Nov/2008:13:21:43 +0100] "OPTIONS * HTTP/1.0" 200 - "-"
> "Apache (internal dummy connection)"
> 127.0.0.1 - - [06/Nov/2008:13:21:45 +0100] "OPTIONS * HTTP/1.0" 200 - "-"
> "Apache (internal dummy connection)"
> 127.0.0.1 - - [06/Nov/2008:13:21:46 +0100] "OPTIONS * HTTP/1.0" 200 - "-"
> "Apache (internal dummy connection)"
> 127.0.0.1 - - [06/Nov/2008:13:21:48 +0100] "OPTIONS * HTTP/1.0" 200 - "-"
> "Apache (internal dummy connection)"
> 127.0.0.1 - - [06/Nov/2008:13:21:49 +0100] "OPTIONS * HTTP/1.0" 200 - "-"
> "Apache (internal dummy connection)"
>
> Note: If I access directly one of the webservers on its real IP, my access
> is correctly logged. This only happens with requests forwarded by LVS.
>
>
> 2008/11/6 Agostino di Salle <ago@xxxxxxxxxxxx>
>
> > never seen this kind of problem with LVS-DR and any kind of web server
> > (apache, IIS, etc)...
> >
> > Which exactly is your setup? can you send an extract of the access_log?
> >
> > Regards,
> > Agostino
> >
> > --
> > Experience is when you get what you don't want to get!
> >
> > On Thu, 6 Nov 2008 at 11:29, Julien Cornuwel wrote:
> >
> > > Hi,
> > >
> > > I've got an LVS-DR load-balancer that works fine.
> > >
> > > The problem is that whatever approach is used (VIP on lo, or iptables
> > > redirect), apache sees the requests as comming from the loopback
> > interface
> > > (which they are). And then, the access_log is filled with "Apache
> > (internal
> > > dummy connection)" instead of real infos on the clients.
> > >
> > > This is obviously an Apache problem but I couldn't find a way to make it
> > > treat requests on loopback and on eth0 the same way. Are you all
> > > experiencing the same issue or did you find a way to make Apache beleive
> > > requests are comming from the outside and log them properly ?
> > >
> > > Regards,
> > > _______________________________________________
> > > LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx
> > > Send requests to lvs-users-request@xxxxxxxxxxxxxxxxxxxxxx
> > > or go to http://lists.graemef.net/mailman/listinfo/lvs-users
> > >
> >
> >
> > _______________________________________________
> > LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx
> > Send requests to lvs-users-request@xxxxxxxxxxxxxxxxxxxxxx
> > or go to http://lists.graemef.net/mailman/listinfo/lvs-users
> >
> _______________________________________________
> LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx
> Send requests to lvs-users-request@xxxxxxxxxxxxxxxxxxxxxx
> or go to http://lists.graemef.net/mailman/listinfo/lvs-users
>
|
