Node object has two ip adresses, causes system alarm about PU410 node down.
I have a reocurring system alarm, which I think is caused because there are two entries for IP adresses in the node object. In attachement the whole setup is displayed.
Anyone know a solution to solve this? I had made a mistake at the initial setup of the system (rnrp config) and this is the result from that mistake.
Voted best answer
A PU410 should be on a local RNRP network area (suggested IP address is 172.16.168.x - third byte contain "128" which hides the area, the remainder is 40 throwing RNRP Area 10) making it invisible to all nodes, except for the ones with physical connection to the area.
If AS and CS is combined, the System Alarm service will see & monitor the local PU410 (but not the redundant which is "beyond" another server that should have Local flag set for the PU410 area).
If the RNRP Monitor (and log file; C:\ProgramData\RnrpEvent.log) agree, i.e. PU410 being reported down there seem to be some issue with the network; check NIC settings, cabling, VMware settings (type of virtual NIC, etc.), etc.
Regardless of hidden or not hidden, the PU410 connection should not be reported down unless you have unplugged or shut the PU410 down.
The IP-addresses displayed in the Node Definition aspect's Configuration tab is not what you might expect.
The aspect works in "three modes":
Mode 1 : The "own" node's Node Definition is called for display (the node where your workplace was launched in): the addresses returned come from "ipconfig" (all local interfaces with an IP address will show up).
Mode 2: A different node is being displayed (your workplace is running at ENG1, but you called AS1's Node Definition): the addresses returned come from what the "hosts" file contain in your node. Starting with 5.1, RNRP injects addresses into the hosts file. In previous versions, I believe a DNS request is made and the content is drawn from the DNS server's Forward Lookup Zone.
Below; I have "injected" a phony address into the hosts file - which is then showed when calling the Node Definition aspect.
Mode 3: The aspect has been manually edited (which is possible after hitting the "F2" key) and the addresses are hardcoded. This *was* recommended in SV 2.1, but should not be made in SV 3 and later; nodename is expected!
So depending on what you checked it can be:
a) A problem with the network that need to be sorted out.
b) You checked your local machine and some additional NIC was listed; no worries...
c) You checked a remote node, but the hosts file (or DNS) need to be cleaned from the bad address
d) Someone has manually edited the entry. Restore original (nodename)
Why have you changed the "Base Adress (implicit adresses)"? If you use the standard entry "172.16.0.0 then your PU410 adresses will be implicit as Network 10. You need only "Number of used exlicit addr Tabs = 1" on both servers for your 152.128... adress.
But I don't know if it solve your original problem, just give it a try. Remember to restart RNRP.
Did you observe the "mode issue" when reviewing IP addresses in the Node Definitions?
Any edited data should be reverted to pure/sole hostname. Then the hosts file content will be shown in mode 2 (viewing another node's addresses).
Compare hosts files if mode 2 show differences - use RNRP wizard to refresh hosts file(s).
The "Register This Connection In DNS" setting on a NIC decide if the node/NIC will be inserted to remote hosts files or not. RNRP Monitor will list a node appended with "h,name=..." if this setting is enabled and only "name=..." if not. The setting should be enabled for NICs on all Client/Server network paths, but *may* be off for other NICs to prevent undesired addresses to show up in the Node Definition.
Last; the RNRP Filter in the Base Config Wizard should be set with the Client/Server network paths to restrict 800xA traffic to those paths only.