Simon, I'm new to this, so if this doesn't help, don't despair - the more experienced members will be along shortly. :-)
Could it be that you need "stickyness?" I think that's the term for the concept you are describing. Also, if that's a two node cluster, have you defined cluster property no-quorum-policy="ignore"? Best regards, Mike On Mon, Dec 13, 2010 at 7:27 AM, Simon Jansen <simon.jans...@googlemail.com>wrote: > Hi, > > I'm trying to set up location contraints for my cluster, but I don't get > them to work in the way that I want. > > The constraints should implement the following behaviour: > - Normal operation > msDRBD0 and resIP0 start on node1, msDRBD1 and resIP1 start on node2 > - Loss of network connection > When the network connection to the public network is lost (tested with > pingd resource) the corresponding DRBD resource should failover to the other > node. > When the connection comes back again the cluster should go into normal > operation again. > > > The pingd resource is configured as follows: > primitive resPing ocf:pacemaker:pingd \ > params host_list="node1 node2 standard-gateway" dampen="5s" interval="2" > multiplier="10000" \ > op start timeout="90" op stop timeout="120" > where node1, node2 and standard-gateway stands for the corresponding IP > adress. > > The definition of the location contraints to bind the DRBD partition to one > node: > location locDRBD0Node1 msDRBD0 \ > rule $id="locDRBD0Node1-rule" $role="Master" 1000: #uname eq node1 > location locDRBD1Node2 msDRBD1 \ > rule $id="locDRBD1Node2-rule" $role="Master" 1000: #uname eq node2 > location locIP0Node1 resIP0 \ > rule $id="locIP0Node1-rule" 1000: #uname eq node1 > location locIP1Node2 resIP1 \ > rule $id="locIP1Node2-rule" 1000: #uname eq node2 > > > I tried the following location constraint: > location locDRBD0Ping msDRBD0 rule $role=Master -inf: not_defined pingd or > pingd lte 10000 > > But with this constraint the cluster does not go into normal operation > again. > > > Maybe someone has implemented an assimilable configuration. > Thanks for any answers. > > -- > > > Regards, > > Simon Jansen > > > --------------------------- > Simon Jansen > 64291 Darmstadt > > _______________________________________________ > Pacemaker mailing list: Pacemaker@oss.clusterlabs.org > http://oss.clusterlabs.org/mailman/listinfo/pacemaker > > Project Home: http://www.clusterlabs.org > Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf > Bugs: > http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker > > -- Mike Diehn Senior Systems Administrator ANSYS, Inc - Lebanon, NH Office mike.di...@ansys.com, (603) 727-5492
_______________________________________________ Pacemaker mailing list: Pacemaker@oss.clusterlabs.org http://oss.clusterlabs.org/mailman/listinfo/pacemaker Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker
