Ethernet - Avaya Media Processing Server 1000 Hardware Installation And Maintenance

Operation and Administration

Ethernet

Each NIC provides a redundant Ethernet interface, (Ethernet A, Ethernet B). The NIC supports two
types of interfaces: an Ethernet hub or an Ethernet switch. The external Ethernet connection is on
the rear panel of the chassis. All base TMS and NIC communication is routed through the Ethernet
hub/switch on the NIC.
After initial power up, the boot ROM on the NICs arbitrate for master/slave, which is performed
before loading the runtime image for the NIC. The master enables Ethernet A, and the slave
enables Ethernet B. In non-redundant mode (that is, single NIC), both ports are enabled by the NIC.
If a failure occurs, the NIC master enables both ports. When the redundant NIC becomes available,
it becomes the slave by default. The master disables Ethernet B and the slave enables Ethernet B.
A software-controlled failover from master to slave does not cause the NICs to change control of the
Ethernet ports. The impact of the failover is minimized and corruption of Ethernet messages is
prevented.
An Ethernet failure has different results depending on the nature of a fault and what state the
system is in when a fault occurs. An Ethernet fault on initial power up of a chassis can prevent any
or all hardware from loading its run-time image. A failure on an operational system is detected as a
failed health check. Note that a failed health check does not necessarily indicate an Ethernet fault.
On initial power up of a chassis, if the power-on arbitration fails the Ethernet ports are not enabled
correctly. Either or both NICs can enable or disable the A port. In either case, the TFTP boot
protocol for all devices is affected, and TMS and NIC cannot load their image.
At run time, an Ethernet fault can affect the entire chassis or an individual piece of hardware.
Because each board has a unique connection, a failure can occur on a single TMS port. Any run-
time failure of the Ethernet is detected through health check failures.
The following is a list of reasons for Ethernet failure:
• faulty Ethernet hardware on NIC
• cabling error
• faulty boot ROM
• server down or incorrect configuration
• IP or MAC address conflicts
• Ethernet hub/switch error
• NIC Arbitration failing
• errors in bootptab
• hub on server is not operating correctly
• poor or intermittent connection with backplane
Troubleshooting
1. Inspect all cabling.
2. Issue the ping command to verify IP/MAC addresses.
Avaya Media Processing Server 1000 Hardware Installation and Maintenance
190
Comments? infodev@avaya.com
October 2014