Coyote Point Systems Equalizer GX Series, Administration Manual

Page: 544 / 866

Failover
GUI Parameter (CLI Parameter) Description
Failed Probe Count
(
strike_count )
The number of successive failed heartbeats that must occur before a peer is
marked "down" (default:3). A heartbeat is considered to have failed
whenever the
Heartbeat Interval has elapsed and no probe has been
received from a peer during that interval.
If the
Failed Probe Count
reaches its specified maximum value on the sub-
net, or if the
Failed Probe Count
reaches 1 on all subnets in a multi-subnet
network configuration, then a failover can occur. If the subnet
Failed Probe
count is 0, then the global
Failed Probe Count (strike_count)
is used
Retry Interval
(
retry_interval )
Time in seconds (default:5) between checks for changes in Equalizer’s con-
figuration, for the purpose of determining whether a configuration transfer
to the remote peer is required.
The Retry Interval is also used when any failover operation other than
heartbeating fails. For example, if the system is rebooted and a VLAN health
check fails, we try again
Connect Timeout
(
conn_timeout
)
Time in seconds (default:1) to wait for a TCP connection to be established
between peers.
Receive Timeout
(
recv_timeout
)
Time in seconds (default:1) after a TCP connection is established, of how
long a peer waits for the other peer’s response.
Failover Occurs If...
l
The number of failed probes on any single subnet equals or exceeds the
Failed Probe Count
for
that subnet.
OR
l
The number of subnets with a
Failed Probe Count
greater than 0 equals or exceeds the global
Failed Probe Count.
Modifying Failover Timeouts in Production
When an failover pair is actively serving traffic, any changes to the global or subnet failover
parameters could result in a failover if you do not perform them in the correct manner.
Note
- These parameters are not currently synchronized.
To prevent a failover from occurring , proceed with the following to prevent an inadvertent
failover:
1. Disable failover on both units by disabling the failover flag on the local peers on both units;
first on the current primary unit and then the current backup. Refer to
"Configuring Act-
ive/Passive Failover (CLI)"
on page 562 or
"Configuring Active/Passive Failover (GUI)"
on page 572
for details.
2. Make the changes to failover timeouts required on both systems, in any order. Refer to
"Fail-
over Peer Probes and Timeouts"
on page 543 if necessary.
3. Re-enable failover by checking the failover flags on the on the primary Equalizer's local
peer and then on the backup.
544 Copyright © 2014 Coyote Point Systems, A Subsidiary of Fortinet, Inc.