Try the documentation at http://clusterlabs.org/wiki/Documentation instead The section on clones might be of particular interest given that you want an active-active apache install. On Fri, Mar 6, 2009 at 17:23, LLLActive@GMX.Net <LLLActive@gmx.net> wrote:
Hi all,
I am busy configuring an active/active HA2 (cib.xml) web server with SLES 10 SP2. I started manually with configuring a very basic with the YaST GUI and then manually checked ha.conf; thereafter I used hb_gui.
For a manual config, I find the http://www.linux-ha.org website documentation a little difficult to follow between HA versions 1 & 2. They do not have a setup only for Version 2; you have to read all V1 and then find the V2 differences like using the cib.xml in stead of the haresources config files. I think I managed to get the basic heartbeat running as far as configuring the ha.conf file manually (and with Yast & hb_gui). I have set up a failover with a common IP and switched with hb_gui, and checked with ssh that the server actually switched. I then addad a Webserver on the common IP, which also works with a manual switching in hb_gui; actually stopping it on one server and starting on the other. Automatic failover does not seem to work.
I found a more systematic description at http://www.novell.com/documentation/sles10/heartbeat/index.html?page=/docume... for heartbeat.
I can find no actual description of setting up an active-active system which runs a web server on each node separately, that fails over to the other when a node fails. It means that at fail over, the one node should run both web servers, switching back to the default server as soon as it comes up again. I created a group which first switches the IP and then the web server. It works with a manual switch over, but not automatically. This is only switching from a resource on one node to another, i.e. another web server. I will create a NFS common data share for both on the same DRBD HA data cluster as a htdocs directory, then the content will be the same source. This is the one web server. Another web server has to host an ERP II (Web 2.0) system with its own database (also on a DRBD HA cluster) on the other node. Whenever one node fails, the other should take over and host both web servers.
What I need then, are two web servers running each on its own node, and the takeover of the failing web server by the other node, which then runs both web servers till the default preferred server for the failed website comes back on.
Any indicators as where to look for such a setup is highly appreciated.
TIA
:-) Al -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org