The Low Latency Transport (LLT) protocol is a layer 2 communication protocol which lives on the Data Link Provider Interface (DLPI). The dlpiping utility is provided with the VRTSllt package to test communication of specific DLPI devices at this level.
A dlpiping server must be configured on one of the NICs to be tested. This server will reflect packets sent by other DLPI devices attached to the network and configured in client mode.
1. On one node in the cluster, perform the following steps to configure a dlpiping server:
Use the getmac command to display the MAC address of the device on which the dlpiping server will be configured. Record this address for use in steps performed on the client systems. The getmac command takes the form of getmac <device>:<instance>
# /opt/VRTSllt/getmac /dev/qfe:0
Configure the dlpiping server on this interface with the dlpiping command. This command takes the form of dlpiping -s <device>:<instance>. Note that this server will run in the foreground (no prompt will be returned). Once testing is completed, use ^C to stop the server process.
# /opt/VRTSllt/dlpiping -s /dev/qfe:0
At this point, you can configure other nodes in the cluster as dlpiping clients.
2. Using the MAC address obtained in step 1 above, perform the following on all nodes to test:
Configure the dlpiping client to send packets to the dlpiping server configured above. If communication is successful, you will receive a message indicating that the MAC address specified is alive:
# /opt/VRTSllt/dlpiping -c /dev/qfe:0 08:00:20:E7:DE:B2
08:00:20:E7:DE:B2 is alive
If communication is NOT successful, you will receive output indicating there was no response from the specified MAC address:
# /opt/VRTSllt/dlpiping -c /dev/hme:1 08:00:20:E7:DE:B2
no response from 08:00:20:E7:DE:B2
If there is no response, there is most likely a problem with your hardware configuration. For example, you may be using the wrong type of cables for your environment (ie, standard ethernet cables as opposed to crossover cables) or you may have the server configured on the wrong Network Interface Card (NIC) (ie, /dev/qfe:0 instead of /dev/qfe:1). Identify possible configuration errors, and repeat the steps above.
Note that the examples above use device interfaces from a Solaris system as an example.
Other platform device naming conventions differ from the Solaris naming conventions.
Example network device naming for other platforms are shown below.
For Linux, interface names are simply "eth0" or "eth1", etc. So an example Linux command could be:
# /opt/VRTSllt/dlpiping -s eth0
For AIX, interface names follow the pattern of "dev/dlpi/(interface)". So an example AIX command could be (for an interface 'en1'):
# /opt/VRTSllt/dlpiping -s /dev/dlpi/en:1