<font face="courier new,monospace">Hi All,</font><div><font face="courier new,monospace"><br></font></div><div><font face="courier new,monospace"><br></font></div><div><font face="courier new,monospace">I am following the high availability guide provided by openstack as seen here: </font><a href="http://docs.openstack.org/trunk/openstack-ha/content/ch-intro.html">http://docs.openstack.org/trunk/openstack-ha/content/ch-intro.html</a></div>
<div><br></div><div>I currently have mysql and rabbitmq setup with pacemaker and corosync. Everything seems to be working fine.</div><div><br></div><div>I am now adding in the openstack services. I have keystone configured and it is using haproxy for load balancing/redundancy and keepalived for the virtual IP. haproxy is set configured to bind to non-local address and is running on all HA nodes. </div>
<div><br></div><div>Is this a good way to make the openstack services such as glance, swift, nova and keystone HA? </div><div><br></div><div>Is it preferable to have keepalived start haproxy on the node that it assigns the virtual IP to, or is my setup with the non-local binding ok?</div>
<div><br></div><div>Is this (<a href="http://docs.openstack.org/folsom/openstack-network/admin/content/ha_pacemaker.html">http://docs.openstack.org/folsom/openstack-network/admin/content/ha_pacemaker.html</a>) the preferred to make the networking (quantum) HA?</div>
<div><br></div><div><br></div><div>and one last question:</div><div><br></div><div>What is a good method to detect a compute node failure and automatically restart the virutal machines it was running on another node?<br><br>
Thank you so much, and sorry to bombard you with questions!</div><div><br></div><div>Sam </div><div><br></div><div><br></div>