Chapter 15 High Availability Troubleshooting
The ATCA environment will usually contain a high-availability failover configuration between
two ATCA switches in the chassis. Note that the failover features are configurable and a switch
can be directed to fail over all of its processing when a single port or link goes down, or it can
perform a port-to-port or VLAN-to-VLAN failover where both partner switches are still
processing a portion of the network traffic.
Before replacing a switch that has gone out of service because of a switch-level failover, you
need to understand how the high-availability features have been configured. If the switch failover
was triggered by a port or link failure, make sure to isolate the cause for the link failure first, to
make sure the problem is not external to the switch (for example, a bad or loose cable for a wired
port).
Spontaneous Failover Activity
If while rebooting the inactive switch in a chassis causes the active switch to reboot and/or an
unexpected failover, you can try setting the
zsp.conf
file
vrrp_msg_rate
to 500.
The
VRRP_msg_rate
is the time in milliseconds between transmissions VRRP messages on the
inter-switch link (ISL). The VRRP protocol requires the absence of three VRRP messages before
concluding that the remote switch has failed. The
msg_rate
must match the
msg_rate
of all
siblings. Anything other than multiples of seconds does not conform to the VRRP specification,
and will only run with the
vrrpd
.
Unexpected Fail-back Activity
If unexpected fail-back activity is observed check to make sure that only one switch is setup as
the Master switch (
vrrpd –M option
) or the switches will oscillate. See the Ethernet Switch
Blade User’s Guide for more information on setting the failover priority level.
Ethernet Switch Blade User's Guide
release 3.2.2j
page 183
Downloaded from
www.Manualslib.com
manuals search engine