I have a 3 node cluster running, volume, and DHCP resources.
On Similar hardware 2X HP DL380 G3, 1X HPDL360 G3. It appears that when
volume resources are failed over to the 360 the address related secondary
ip address can only be seen within the local VLAN. When the DHCP resource
is failed to this server it can be seen accross all networks. This is the
case for when all volume resources are failed to this server, even though
the Secondary address is succesfully binding and there is no error.
Am looking at the switch (Cisco 4507 Layer3) as a cause at this stage but
am open to any suggestions. Also am hesitant to alter ARP settings on our
core switch, when no other servers are having problems with ARP refresh.
Another difference I can see in setup is the "IP-Address helper"(Bootp
forwarder) assigned to DHCP resource IP however this is refreshing fine
when switching between other servers. Have thought maybe NIC is not
advertising to switch correctly after secondary address binds ?
Any assistance or sugestions will be greatly appreciated.