RHEL5 や RHEL6 上において master_wins モードで qdisk を使用し、1 つのノードが不意にクラスタを離脱する時、rgmanager が "Quorum Dissolved" を報告して全てのリソースを停止します

Issue

• RHEL5 や RHEL6 上において master_wins モードで qdisk を使用している時、クオーラムが解消されました。

Apr  1 15:25:55 rh6node2 corosync[14188]:   [TOTEM ] A processor failed, forming new configuration.
Apr  1 15:25:57 rh6node2 kernel: dlm: closing connection to node 1
Apr  1 15:25:57 rh6node2 corosync[14188]:   [CMAN  ] quorum lost, blocking activity
Apr  1 15:25:57 rh6node2 corosync[14188]:   [QUORUM] This node is within the non-primary component and will NOT provide any services.
Apr  1 15:25:57 rh6node2 corosync[14188]:   [QUORUM] Members[1]:2
Apr  1 15:25:57 rh6node2 corosync[14188]:   [TOTEM ] A processor joined or left the membership and a new membership was formed.
Apr  1 15:25:57 rh6node2 corosync[14188]:   [CPG   ] chosen downlist: sender r(0) ip(192.168.122.162) ; members(old:2 left:1)
Apr  1 15:25:57 rh6node2 corosync[14188]:   [MAIN  ] Completed service synchronization, ready to provide service.
Apr  1 15:25:57 rh6node2 rgmanager[14577]:#1:Quorum Dissolved
Apr  1 15:25:57 rh6node2 rgmanager[27697]:[ip] Removing IPv4 address 192.168.122.222/24 from eth1
Apr  1 15:25:57 rh6node2 avahi-daemon[2018]:Withdrawing address record for 192.168.122.222 on eth1.
Apr  1 15:25:57 rh6node2 rgmanager[27733]:[script] Executing /etc/cluster/scripts/test_script-1.sh stop
Apr  1 15:25:57 rh6node2 rgmanager[27755]:Executing stop on the script:/etc/cluster/scripts/test_script-1.sh
Apr  1 15:25:57 rh6node2 rgmanager[27773]:The script is not running:/etc/cluster/scripts/test_script-1.sh.
Apr  1 15:26:03 rh6node2 corosync[14188]:   [CMAN  ] quorum device re-registered
Apr  1 15:26:03 rh6node2 corosync[14188]:   [CMAN  ] quorum regained, resuming activity
Apr  1 15:26:03 rh6node2 corosync[14188]:   [QUORUM] This node is within the primary component and will provide service.
Apr  1 15:26:03 rh6node2 corosync[14188]:   [QUORUM] Members[1]:2
Apr  1 15:26:03 rh6node2 qdiskd[14239]:Assuming master role
Apr  1 15:26:03 rh6node2 fenced[14409]: fencing node rh6node1.examplerh.com
Apr  1 15:26:03 rh6node2 qdiskd[14239]:Writing eviction notice for node 1
Apr  1 15:26:04 rh6node2 qdiskd[14239]:Node 1 evicted
Apr  1 15:26:05 rh6node2 fenced[14409]: fence rh6node1.examplerh.com success
Apr  1 15:26:07 rh6node2 rgmanager[27915]:[ip] Removing IPv4 address 192.168.122.221/24 from eth1
Apr  1 15:26:17 rh6node2 rgmanager[27975]:[ip] Removing IPv4 address 192.168.122.220/24 from eth1
Apr  1 15:26:27 rh6node2 rgmanager[14577]:Quorum Regained

• master_wins モードで qdisk を使用しているクラスターノードの 1 つが失敗した時、全てのクラスター化されたサービスとリソースが停止されました。