2 Replies Latest reply on Feb 28, 2013 10:07 AM by RichardLetts

    Atlas Map - Two Cisco Switch Connected to Multiple NICs in Servers


      In Atlas Map I'm attempting to map out the SAN environment to include two switches, the 2 SAN controllers (8 eths), and the multiple NICs on the servers.  However, when I add the two switches and the servers the connect now doesn't work.  I went into the properties of the servers and dragged the iSCSI interface, however, it then reports the node isn't managed. These switches are in a separate VLAN.  The node is managed with the resource monitored.  Is this task even possible without exporting the MAC addresses from switch then changing the MAC to IP and add all interfaces and switch ports manually?


      The disk performance is vital in the environment and so with a diagram and drill-downs to the active iSCSI server interface will help determine bottlenecks.  



        • Re: Atlas Map - Two Cisco Switch Connected to Multiple NICs in Servers

          Using LANSurveyor I was able to determine the ports on the cisco switch with the particular interface(s) of the node.  Since Atlas Map does not automatically connect interfaces on Window Servers to the port on the switch it makes Atlas Map extremely time consuming.  Furthermore, these manual connections in AtLas Map only reports if it is up or down if the "Connect Now" does not work.


          In addition, in Atlas Map, if a node connects to two switches before connecting to a the core switch, the "Connect Now" does not connect the lines from node --> switch 4  --> switch 3 --> switch 1.  Instead, it connects from node --> switch 1.  Not cool.


          Please implement LANsurveyor to replace AtLas Map.

          • Re: Atlas Map - Two Cisco Switch Connected to Multiple NICs in Servers

            are you gathering layer-3 information from the switches? if so, try turning it off may help -- SWO is using the IP<>ARP tables for connection rather than the bridging tables.


            I had the same problem with routers and switches across our northernwave network where internal management interfaces were being shown as connected to adjacent switches instead of the actual physical interfaces. turning off gathering L3 information from the switches helped 'fix' the autoconnect.


            as an aside: I think that reporting bugs works better opening tickets with SW using your support contract.