Can't ping Nimble Data Discovery IP across VLANs

[Flames89]

We just completed a core switch replacement, moving from Cisco Catalyst to Aruba 3810M switches, and since then the Nimble data discovery IP is only reachable from within the same VLAN as it (10.0.0.0). The management interface is on a different VLAN (10.0.2.0) and I can reach it from every other subnet without issue. Our ESX ISCSI interfaces are in the 10.0.0.0 subnet so there's no issue with VM storage, however Nimble replication to our other office is broken as the remote Nimble is on a different subnet. There's no issue on the switch that I can see, every other device on the 10.0.0.0 subnet can be reached from all other VLANs, it's only the Nimble Data IP that can't. Did I miss something that's specific to Nimble when setting up the new switch, or is there something else wrong? 

[swilson120]

Hi, could be many things.

We use Extreme Network switching and you have to enable the vlan for IP forwarding.  I don't know if Aruba has the same requirement or not.

Also, we replicate from our production to site to DR site using the management interfaces, not the data interfaces.  If you can ping between those two, setup your replication to use those interfaces.

Lastly, on the production SAN there might be a command to look at the route and arp tables.  Make sure your default gateway is the IP address of your core switch on the mgmt interface.   On linux you can do a "route print" command to see if there is a static route defined for your DR site in the route table that is using a non existing gateway.  Also, make sure the arp address of core switch is displayed on the SAN.  If you have a duplicate IP address issue, the SAN will have the wrong arp address of the core switch.  

There are things I would look at.

Scott