wilsj
asked on
Failover using rgmanager in RHCS5
Hey all,
I have a cluster up and running with gfs, drbd, and gnbd. I also have drbd running primary/primary. I am trying to setup a failover ip for the 2 drbd servers so if one goes down the exported gnbd devices are still there. I set up the failover IP through RHCS5 using rgmanager but for some reason when I pull the plug on server1 the ip address does automatically go over to server2. I'm thinking maybe I didn't set up the resources correctly. any help is appreciated.
I have a cluster up and running with gfs, drbd, and gnbd. I also have drbd running primary/primary. I am trying to setup a failover ip for the 2 drbd servers so if one goes down the exported gnbd devices are still there. I set up the failover IP through RHCS5 using rgmanager but for some reason when I pull the plug on server1 the ip address does automatically go over to server2. I'm thinking maybe I didn't set up the resources correctly. any help is appreciated.
Can you post here you /etc/cluster/cluster.conf file?
ASKER
sure no problem. and thanks for the reply.
Dolphins is my storage, drbd, gfs, gnbd server
Patriots is replicating the data on Dolphins with drbd
lions is my xen server and imports the storage from Dolphins/Patriots.
The idea I had was is dolphins fails then just the ip needs to switch to patriots and the same gnbd devices will already be exported and the xen instances won't lose connection.
<?xml version="1.0"?>
<cluster config_version="9" name="nas-cluster">
<fence_daemon post_fail_delay="0" post_join_delay="3"/>
<clusternodes>
<clusternode name="dolphins" nodeid="1" votes="1">
<fence>
<method name="1">
<device name="nas-fencing" nodename="dolphins"/>
</method>
</fence>
</clusternode>
<clusternode name="lions" nodeid="2" votes="1">
<fence>
<method name="1">
<device name="nas-gnbd-fencing" nodename="lions"/>
</method>
</fence>
</clusternode>
<clusternode name="patriots" nodeid="3" votes="1">
<fence>
<method name="1">
<device name="nas-fencing" nodename="patriots"/>
</method>
</fence>
</clusternode>
</clusternodes>
<cman/>
<fencedevices>
<fencedevice agent="fence_gnbd" name="nas-gnbd-fencing" servers="dolphins-storage- failover"/ >
<fencedevice agent="fence_manual" name="nas-fencing"/>
</fencedevices>
<rm>
<failoverdomains>
<failoverdomain name="dolphins-drbd1" ordered="1" restricted="0">
<failoverdomainnode name="dolphins" priority="1"/>
<failoverdomainnode name="patriots" priority="2"/>
</failoverdomain>
</failoverdomains>
<resources>
<ip address="192.168.5.4" monitor_link="1"/>
</resources>
<service autostart="1" domain="dolphins-drbd1" name="dolphins-svc-drbd1" recovery="relocate">
<ip ref="192.168.5.4"/>
</service>
</rm>
<fence_xvmd/>
<fence_xvmd/>
</cluster>
Dolphins is my storage, drbd, gfs, gnbd server
Patriots is replicating the data on Dolphins with drbd
lions is my xen server and imports the storage from Dolphins/Patriots.
The idea I had was is dolphins fails then just the ip needs to switch to patriots and the same gnbd devices will already be exported and the xen instances won't lose connection.
<?xml version="1.0"?>
<cluster config_version="9" name="nas-cluster">
<fence_daemon post_fail_delay="0" post_join_delay="3"/>
<clusternodes>
<clusternode name="dolphins" nodeid="1" votes="1">
<fence>
<method name="1">
<device name="nas-fencing" nodename="dolphins"/>
</method>
</fence>
</clusternode>
<clusternode name="lions" nodeid="2" votes="1">
<fence>
<method name="1">
<device name="nas-gnbd-fencing" nodename="lions"/>
</method>
</fence>
</clusternode>
<clusternode name="patriots" nodeid="3" votes="1">
<fence>
<method name="1">
<device name="nas-fencing" nodename="patriots"/>
</method>
</fence>
</clusternode>
</clusternodes>
<cman/>
<fencedevices>
<fencedevice agent="fence_gnbd" name="nas-gnbd-fencing" servers="dolphins-storage-
<fencedevice agent="fence_manual" name="nas-fencing"/>
</fencedevices>
<rm>
<failoverdomains>
<failoverdomain name="dolphins-drbd1" ordered="1" restricted="0">
<failoverdomainnode name="dolphins" priority="1"/>
<failoverdomainnode name="patriots" priority="2"/>
</failoverdomain>
</failoverdomains>
<resources>
<ip address="192.168.5.4" monitor_link="1"/>
</resources>
<service autostart="1" domain="dolphins-drbd1" name="dolphins-svc-drbd1" recovery="relocate">
<ip ref="192.168.5.4"/>
</service>
</rm>
<fence_xvmd/>
<fence_xvmd/>
</cluster>
ASKER
Am I supposed to create a resource for the partition that is being shared? If yes this is confusing for me. Because I already have the partition running on the other server patriots. All I want is for the ip to switch servers is this possible with red hat cluster suite 5? Thanks again.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
i finally figured it out thanks for the replies