Hi Dimitre, VIP on node where public eth is down, will failover to other node/nodes and all resources dependant on it will also failover ... dependant resources you can find with following: oracle@main-gc-1@+ASM1: ~ -bash-3.2 $ crsctl stat res -p | egrep -i '^name|START_DEPENDENCIES' NAME=ora.GC_OCR_VOTE.dg START_DEPENDENCIES=hard(ora.asm) pullup(ora.asm) NAME=ora.LISTENER.lsnr START_DEPENDENCIES=hard(type:ora.cluster_vip_net1.type) pullup(type:ora.cluster_vip_net1.type) NAME=ora.LISTENER_SCAN1.lsnr START_DEPENDENCIES=hard(ora.scan1.vip) dispersion:active(type:ora.scan_listener.type) pullup(ora.scan1.vip) NAME=ora.LISTENER_SCAN2.lsnr START_DEPENDENCIES=hard(ora.scan2.vip) dispersion:active(type:ora.scan_listener.type) pullup(ora.scan2.vip) NAME=ora.LISTENER_SCAN3.lsnr START_DEPENDENCIES=hard(ora.scan3.vip) dispersion:active(type:ora.scan_listener.type) pullup(ora.scan3.vip) NAME=ora.asm START_DEPENDENCIES=weak(ora.LISTENER.lsnr) NAME=ora.cvu START_DEPENDENCIES=hard(ora.net1.network) pullup(ora.net1.network) NAME=ora.gsd START_DEPENDENCIES= NAME=ora.main-gc-1.vip START_DEPENDENCIES=hard(ora.net1.network) pullup(ora.net1.network) NAME=ora.main-gc-2.vip START_DEPENDENCIES=hard(ora.net1.network) pullup(ora.net1.network) NAME=ora.net1.network START_DEPENDENCIES= NAME=ora.oc4j START_DEPENDENCIES= NAME=ora.ons START_DEPENDENCIES=hard(ora.net1.network) pullup(ora.net1.network) NAME=ora.registry.acfs START_DEPENDENCIES=hard(ora.asm) pullup:always(ora.asm) NAME=ora.scan1.vip START_DEPENDENCIES=hard(ora.net1.network) dispersion:active(type:ora.scan_vip.type) pullup(global:ora.net1.network) NAME=ora.scan2.vip START_DEPENDENCIES=hard(ora.net1.network) dispersion:active(type:ora.scan_vip.type) pullup(global:ora.net1.network) NAME=ora.scan3.vip START_DEPENDENCIES=hard(ora.net1.network) dispersion:active(type:ora.scan_vip.type) pullup(global:ora.net1.network) regards, goran On Sun, Mar 4, 2012 at 2:52 PM, Radoulov, Dimitre <cichomitiko@xxxxxxxxx>wrote: > > Hi Goran, > we're running similar tests right now :) > I'm searching the documentation but I don't seem > to find where Oracle describes the correct (expected) behavior > of the processes after a failure. > > I mean - what's expected to happen _automatically_ with the > processes/services > after the failure has been fixed (timeouts, other important details?), > and what needs to be done manually. > > Best regards > Dimitre > > > > On 04/03/2012 11:52, goran bogdanovic wrote: > >> small clarification of term "surviving node" ... the node is "surviving" >> in >> term of accessibility from outside of cluster whereas from cluster point >> of >> view, the node is still part of cluster >> On Mon, Feb 6, 2012 at 3:36 PM, Purav Chovatia<puravc@xxxxxxxxx> wrote: >> >> >> > -- //www.freelists.org/webpage/oracle-l