On 9/26/07, Raoul Bhatia [IPAX] <[EMAIL PROTECTED]> wrote: > Dejan Muhamedagic wrote: > > Hi, > > > > On Wed, Sep 26, 2007 at 07:23:43AM -0500, Dave Augustus wrote: > >> On Wed, 2007-09-26 at 00:33 +0200, Dejan Muhamedagic wrote: > >> > >>> Hi, > >>> > >>> On Tue, Sep 25, 2007 at 05:04:16PM -0500, Dave Augustus wrote: > >>>> 6 servers- all running heartbeat > >>>> 1 is LVS loadbalancer > >>>> the other 5 are lvs realservers > >>>> > >>>> 2 resource groups: > >>>> 1 for the loadbalancer > >>>> 1 CLONE resource group for the real servers > >>>> > >>>> The loadbalancer can be ANY of the 6 servers. > >>>> If a server is NOT the load balancer, then it runs as a real server. > >>>> > >>>> Is this a supported configuration? > >>> Yes. If I understood correctly. Just setup a clone with > >>> clone_node_max set to 1. And an extra constraint to make room for > >>> the other resource which would say "don't run these two on the > >>> same node". Furthermore, make sure that the score for the > >>> loadbalancer resource is always higher than for the clone. > >>> > >> > >> OK, I setup currently has 4 nodes. I clone_node_max to 1 and it ran a > >> resource group *twice* on the same node. > >> > >> Is this a bug? > > > > That shouldn't have happened. Please post the configuration and > > logs. > > i think this might be related to my report earlier this month and the > changeset: http://hg.beekhof.net/lha/crm-dev/rev/b570d7658acc
this is definitely the case > so using the current interim build will resolve this issue. nod > otherwise, you can use a constraint to not allow clone:0 and clone:1 on > the same host. _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
