I have a customer who’s trying to add a cluster to OCUM v5.1. This cluster has 3 different cluster management LIFs, only one of which is routable to / from the OCUM Core server. When we add the cluster to OCUM by specifying the routable interface (172.26.128.xx), the cluster and all elements are discovered, but then OCUM flips the 'Cluster IP' it is using to one of the non-routable cluster management LIFs (172.28.128.xx) and we lose connectivity to the cluster. How does OCUM Core select the cluster IP it uses? Is there a way to force it to use a specific IP for a cluster? Does OCUM have any built-in affinity to a LIF on the sam esubnet (the OCUM server has multiple IPs, one of which is on a private 172.28.128.0 subnet that is totally separate from the 172.28.128.0 subnet the 2nd cluster management interface is on)?
Here are the cluster's cluster management LIFs...we want to use xxxxxCLU01_MGMT, but it keeps flipping to xxxxxVSE01_SCRIPT_ACCESS:
GTCNACLU01::> net int show
(network interface show)
Logical Status Network Current Current Is
Vserver Interface Admin/Oper Address/Mask Node Port Home
----------- ---------- ---------- ------------------ ------------- ------- ----
up/up 172.26.128.39/21 xxxxxCLU01-01 e0a-4052
up/up 172.28.128.28/24 xxxxxCLU01-02 e1b-2070
up/up 172.20.0.87/16 xxxxxCLU01-01 e1b-4093
What does your dfm host get <cluster-name> return esp in the hostPrimaryIPAddress field ?
There was a logic built in DFM to go & pick the clus_mgmt IP automatically from cluster.This was done to handle the clusters which were added using the node mgmt IP and not the cluster_mgmt lif. Looks like this seems to have a side effect.
Maybe the non-routable IP shows up 1st in the list of clus_mgmt ZAPI o/p and that gets swapped with the routable IP.
Can you pls open a case with NGS ?