高可用OpenStack理论分析( by quqi99)

作者：张华  发表于：2014-03-05
版权声明：可以任意转载，转载时请务必以超链接形式标明文章原始出处和作者信息及本版权声明

http://blog.csdn.net/quqi99 )

1, DB,  active-passive (pacemarker + DRBD); active-active (Galera), http://blog.csdn.net/quqi99/article/details/9392789

    DB, Percona XtraDB Cluster

2, MQ, mq cluster, http://blog.csdn.net/quqi99/article/details/9394121

3, Cinder,

    Local File System (Raid10 + LVM),  http://blog.csdn.net/quqi99/article/details/9396413

    Distrubte File System (Ceph), http://blog.csdn.net/quqi99/article/details/10894833

4, All stateless services, like (keystone|glance|nova|neutron)-api, nova-schedule etc (haproxy + pacemarker)

5, L3-agent (one statefull service), VRRP + keeplived + ip conntracked, http://blog.csdn.net/quqi99/article/details/18799877