[Linux-HA] Failover-Cluster: 2 nodes in different locations

Egon Reetz e.reetz at hamburg.de
Tue Jul 19 05:27:22 MDT 2005


Am 19.07.2005 um 12:29 schrieb Oliver Baltz:

> Hi,
>
> we want a failover-cluster with 2 nodes for a "BSCW"-service
> (http://bscw.gmd.de) connected via 100mBit-LAN.
> That seems to be quite easy with heartbeat/DRBD.
> BUT: Between the 2 nodes there a few routers and switches - it's a  
> LAN of
> an university.
>
> The problem:
> How do the nodes know, whether the node or the network between them is
> down? Only a node-failure should cause a failover-action!
>

Use 2 independent heartbeat connections. Independent means if one  
line goes down, a second line is up
for communication.
Use ucast, ping or ping_group (other nodes normally visible from both  
nodes) within your ha.cf
and have a look what the faq and documentations tells about stonith  
devices.

Egon


> Do you have any solution for me?
>
> Thanks in advance.
> Oliver
>
>
>
>
>
>
> _______________________________________________
> Linux-HA mailing list
> Linux-HA at lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
>




More information about the Linux-HA mailing list