Table 4-6 Community Address ARP Echo Node Probe REPLY Frame Format (continued)
1 byte0x04Protocol address length
2 bytes0x002 (ARP response)ARP operation
6 bytes“Echo Node’s MAC Address”Source MAC address
4 bytes“Echo Node’s IP Address”Source IP address
6 bytes“Community Probe MAC from GUI”Destination MAC address
4 bytes“Community Probe IP from GUI”Destination IP address
4 bytes“varies”Checksum
Community Address ARP’s Designated Echo Node:
An Echo Node is any network device with an IP address that is “designated” by the
implementer for use by a team (with Active Path’s Community Address ARP mechanism
enabled) to validate network connectivity on a per teamed port basis. No special software
or configuration should be required for any device to function as the Echo Node. The
designated Echo Node simply needs to have an IP address and be physically located on the
same broadcast domain as the teamed ports.
If the Echo Node is disconnected from the network or fails to respond to Echo Node probes
for any reason, all teamed ports using the device as an Echo Node will not be able to receive
Echo Node probe replies. When this happens, the teaming driver WILL NOT fail all teamed
ports (in other words, a failed Echo Node will not cause loss of connectivity for the server).
From the team’s perspective, all teamed ports have become equal in regards to Active Path
– none can reach the Echo Node. As a result, the teaming driver will indicate that the teamed
ports are all in the Active Path degraded state (degraded indicates that the port is in use by
the team but isn’t functioning at its optimum level) and are still eligible to be used by the
team. Any teamed port that was previously in the Active Path Failed state due to lack of
communication with the failed Echo Node will also be eligible for use and will be reinstated
to the Active Path degraded state, equal to all other teamed ports in that team. Operationally,
the team will continue to work as if Active Path was disabled on the team since the Echo
Node is not available to any teamed port and the Active Path mechanism can no longer be
used to select teamed ports with better network connectivity. However, Active Path will
continue to send Echo Node probe requests in case the designated Echo Node is restored.
Future implementations of this mechanism may have an option for a backup Echo Node.
While the same Echo Node can be used for every server in the broadcast domain, it is not
required. It is possible to designate a different Echo Node for every Active Path team on
every server on the network. With Community Address ARP, however, the most efficient
implementation is to use a single Echo Node for all servers in the same broadcast domain.
Because the Echo Node probe replies from the Echo Node are received by all teamed ports
configured with the same Community Probe IP/MAC address and same Echo Node, an
Echo Node reply generated from one server’s request can be used to validate teamed ports
for another server’s teamed ports.
HP recommends choosing a designated Echo Node that is physically located on the most
important physical segment of the broadcast domain. Since Active Path validates end-to-end
network connectivity from the teamed port to the Echo Node, the farther the Echo Node is
physically from the server the more network that is validated by Active Path. HP also
recommends choosing an Echo Node that has reliable connectivity to the network. Whenever
the Echo Node is disconnected from the network, Active Path is effectively disabled on any
server configured to monitor the disconnected Echo Node. As mentioned above, a
disconnected Echo Node will not cause the team to disable all teamed ports.
44 The Mechanics of Teaming for the Advanced User
To Next Page
Loading...
Need help?
General