Would it be easier to create a script to bing up the virtual ip and have
rgmanager start the script on failover?
James Wilson wrote:
I removed the second entry for the IP resource but the IP still
doesn't failover to the second node. Here is the output from rg_test
Lon Hohberger wrote:Running in test mode.
Loaded 10 resource rules
=== Resources List ===
Resource type: ip
Instances: 1/1
Agent: ip.sh
Attributes:
address = 192.168.5.4 [ primary unique ]
monitor_link = 1
nfslock [ inherit("service%nfslock") ]
Resource type: service [INLINE]
Instances: 1/1
Agent: service.sh
Attributes:
name = dolphins-svc-drbd1 [ primary unique required ]
domain = dolphins-drbd1
autostart = 1
recovery = relocate
=== Resource Tree ===
service {
name = "dolphins-svc-drbd1";
domain = "dolphins-drbd1";
autostart = "1";
recovery = "relocate";
ip {
address = "192.168.5.4";
monitor_link = "1";
nfslock = "(null)";
}
}
=== Failover Domains ===
Failover domain: dolphins-drbd1
Flags: Ordered
Node dolphins (priority 1)
Node patriots (priority 2)
On Thu, Jul 26, 2007 at 02:30:05PM -0500, James Wilson wrote:
Can someone point why my virtual ip is not failing over. Any help is
appreciated.
<resources>
<ip address="192.168.5.4" monitor_link="1"/>
</resources>
<service autostart="1" domain="dolphins-drbd1"
name="dolphins-svc-drbd1" recovery="restart">
<ip ref="192.168.5.4"/>
</service>
<service autostart="1" domain="dolphins-drbd2"
name="dolphins-svc-drbd2" recovery="restart">
<ip ref="192.168.5.4"/>
</service>
You can't reference an IP in multiple services; the second reference is
ignored. See:
rg_test test /etc/cluster/cluster.conf &> /tmp/foo.out
less /tmp/foo.out
The services should be independent and able to coexist. If they are not
cohabitable, you need to use restricted domains to prevent them from
coexisting.
-- Lon
--
Linux-cluster mailing list
Linux-cluster@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/linux-cluster
--
Linux-cluster mailing list
Linux-cluster@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/linux-cluster
