On 08/17/2010 05:42 AM, Paul Querna wrote: > On Mon, Aug 16, 2010 at 2:28 PM, Andrew Oliver <[email protected]> wrote: >> Note that it is an option, not a default setting. The problem with >> the heartbeat bit which Red Hat/JBoss use is the unstandardized >> proprietary protocol required (http://jboss.org/mod_cluster) with >> separate logic to manage it. > > No, in 2.3/trunk, there is a bundled module to provide this in httpd: > https://svn.apache.org/repos/asf/httpd/httpd/trunk/modules/cluster/mod_heartmonitor.c > https://svn.apache.org/repos/asf/httpd/httpd/trunk/modules/proxy/balancers/mod_lbmethod_heartbeat.c
There is even the code in Tomcat to work with the mod_lbmethod_heartbeat balancer provider see https://svn.apache.org/repos/asf/tomcat/trunk/java/org/apache/catalina/ha/backend. Additionally I added code in trunk (long ago) so Paul's stuff doesn't requires multicast. Trying from the front-end to guess the state of the back-end is tricky, basically all the tricks we try with AJP or this "ping" for http proxy will just check the socket and/or the availability of the back-end.... Well that is already a lot. Speaking about mod_cluster: If there is enough interest I can donate the C part of mod_cluster to httpd (after porting it to trunk) and contribute a listener in TC7 and kind of heartbeat for httpd backend to allow to experiment the concepts. Cheers Jean-Frederic
