[Linux-cluster] dont relocate service after node begin fenced

[Juliano Rodrigues]

While testing RHEL 5.1 to suit my needs (simple samba 2 nodes  
failover) Im running into this issue.

When poweroff node that service is running cluster dont relocate  
service to another node.

If killall smbd, cluster detects that samba dies and relocate properly  
to another node.

Conga display Unknown node state for that node poweroff, but still  
displaying that service is active/running in this node (powered off  
node)???

Cheers, Juliano.


Feb 18 15:17:05 node2 openais[1982]: [CLM  ] CLM CONFIGURATION CHANGE
Feb 18 15:17:05 node2 openais[1982]: [CLM  ] New Configuration:
Feb 18 15:17:05 node2 openais[1982]: [CLM  ]    r(0) ip(10.0.0.2)
Feb 18 15:17:05 node2 openais[1982]: [CLM  ] Members Left:
Feb 18 15:17:05 node2 openais[1982]: [CLM  ] Members Joined:
Feb 18 15:17:05 node2 openais[1982]: [SYNC ] This node is within the  
primary component and will provide service.
Feb 18 15:17:05 node2 fence_manual: Node node1.cluster needs to be  
reset before recovery can procede.  Waiting for node1.cluster to  
rejoin the cluster or for manual acknowledgement that it has been  
reset (i.e. fence_ack_manual -n node1.cluster)
Feb 18 15:17:05 node2 openais[1982]: [TOTEM] entering OPERATIONAL state.
Feb 18 15:17:05 node2 openais[1982]: [CLM  ] got nodejoin message  
10.0.0.2
Feb 18 15:17:05 node2 openais[1982]: [CPG  ] got joinlist message from  
node 2


<?xml version="1.0"?>
<cluster alias="CAMPIXCLU" config_version="22" name="CAMPIXCLU">
         <fence_daemon clean_start="0" post_fail_delay="0"  
post_join_delay="3"/>
         <clusternodes>
                 <clusternode name="node1.cluster" nodeid="1" votes="1">
                         <fence>
                                 <method name="1">
                                         <device name="node1"  
nodename="node1.cluster"/>
                                 </method>
                         </fence>
                 </clusternode>
                 <clusternode name="node2.cluster" nodeid="2" votes="1">
                         <fence>
                                 <method name="1">
                                         <device name="node2"  
nodename="node2.cluster"/>
                                 </method>
                         </fence>
                 </clusternode>
         </clusternodes>
         <cman expected_votes="1" two_node="1"/>
         <fencedevices>
                 <fencedevice agent="fence_manual" name="node1"/>
                 <fencedevice agent="fence_manual" name="node2"/>
         </fencedevices>
         <rm>
                 <failoverdomains>
                         <failoverdomain name="sambaha" ordered="1"  
restricted="1">
                                 <failoverdomainnode  
name="node1.cluster" priority="2"/>
                                 <failoverdomainnode  
name="node2.cluster" priority="1"/>
                         </failoverdomain>
                 </failoverdomains>
                 <resources/>
                 <service autostart="1" domain="sambaha" exclusive="0"  
name="samba00" recovery="relocate">
                         <smb name="vendas" workgroup="CAMPIX">
                                 <ip address="10.0.0.10/24"  
monitor_link="1"/>
                         </smb>
                 </service>
         </rm>
</cluster>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20080218/fc4f21b1/attachment.htm>