Hi, Rodrigo Lima： Have you create the failover segment and add the hosts into the segment? What is the ‘enabled’flag of the segment? 发件人: Sam Su (苏正伟) 发送时间: 2021年12月5日 16:05 收件人: 'rodrigo.lima@o2sistemas.com' <rodrigo.lima@o2sistemas.com> 抄送: 'openstack-discuss@lists.openstack.org' <openstack-discuss@lists.openstack.org> 主题: 答复: [lists.openstack.org代发][kolla-ansible] Default masakari configuration should evacuate instances across nodes in wallaby? Hi, Rodrigo Lima： Have you found the failure notification in the masakari dashboard or DB( Table notifications)? If so, the failure notification would trigger the recovery workflow, and you need to check the masakari-api and masakari-engine log to find the detail. If not, it could be a deployment problem, and you need to check the connectivity between masakari-api and masakari-hostmonitor. 发件人: Rodrigo Lima [mailto:rodrigo.lima@o2sistemas.com] 发送时间: 2021年12月4日 2:56 收件人: openstack-discuss@lists.openstack.org <mailto:openstack-discuss@lists.openstack.org> 主题: [lists.openstack.org代发][kolla-ansible] Default masakari configuration should evacuate instances across nodes in wallaby? I'm working on upgrading an openstack farm from victoria to wallaby. After successful upgrade, I would like to enable hacluster and masakari to test HA between failing compute nodes. Everything seems to be running (pacemaker with the remote nodes OK, corosync without errors, masakari-monitors detects the 2 compute nodes online), but... when I simulate the failure of a node with shutdown, the failure and notification appears in the hostmonitor log, but the instances that were on the failed node don't evacuate, and I couldn't find documentation that explains how to do this specific configuration, if necessary. Does anyone have any ideas?
