Hey guys!
Before opening a bug report, I wanted to share my observations to make sure there is no misunderstanding at my part.
We have a testing environment with two routers (router1
and router2
) with a bunch of VRRP interfaces. After updating from 1.3.2 LTS
to 1.3.3 LTS
I was testing the global startup-delay for VRRP. The config for router1
looks like this:
# show high-availability vrrp
global-parameters {
startup-delay 30
}
group adm {
address 10.236.250.65/27 {
}
hello-source-address 10.236.250.66
interface eth10
peer-address 10.236.250.67
preempt-delay 10
priority 200
vrid 65
}
router2
has the exact same configuration except for a lower priority. While VRRP in general is still working just fine, I noticed that when I reboot router1
, it takes a very long time (about 8 minutes) for preemption to kick in and setting router1
as the master again.
I started playing around with different values for startup-delay
and preempt-delay
and noticed some behaviour I do not quite understand.
Here is a summary:
- with
startup-delay 30
and nopreempt-delay
configured, preemption takes way longer than expected - in order for preemption to work with the configured
preempt-delay
,startup-delay
needs to be configured with a lower value thanpreempt-delay
- a relatively high value for
startup-delay
like120
prevents preemption entirely (at least in the 20 minute timeframe I observed)
The same is true for 1.4-rolling-202307110317
btw.
Am I missing here something or is this a bug?