NimbostratusWhy LTM uses 0.0.0.2 as source IP for monitoring ?
What I did:
create vlan
create route domain
net route-domain 1 {
description f5-inside
vlans {
f5-inside
}
create self IP and default route
net self 10.15.250.6%1/30 {
allow-service default
vlan f5-inside
}
net route default%1 {
gw 10.15.250.5%1
}
now BigIP accessible from my network
Then add nodes and create pool.
ltm node 10.15.3.11%1 {
dynamic-ratio 12
monitor snmp
screen db5
session monitor-enabled
status up
}
ltm node 10.15.3.12%1 {
monitor none
screen db6
session monitor-enabled
}
ltm pool dbs {
members {
10.15.3.11%1:ncube-lm {
session monitor-enabled
}
10.15.3.12%1:ncube-lm {
session monitor-enabled
}
}
monitor tcp
service-down-action reselect
}
All nodes with simple tcp health monitoring became red.
Turn on tcpdump on interface and here what I see:
tcpdump -nn -i f5-inside
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on f5-inside, link-type EN10MB (Ethernet), capture size 108 bytes
12:24:37.193959 IP 0.0.0.2.13452 > 10.15.3.11.161: C=new GetRequest(30) .1.3.6.1.4.1.2021.4.5.0
12:24:43.285081 IP 0.0.0.2.17429 > 10.15.3.12.1521: S 4257356921:4257356921(0) win 5760
12:24:45.386729 IP 0.0.0.2.45812 > 10.15.3.11.1521: S 3199704795:3199704795(0) win 5760
12:24:46.284817 IP 0.0.0.2.17429 > 10.15.3.12.1521: S 4257356921:4257356921(0) win 5760
12:24:47.193349 IP 0.0.0.2.46596 > 10.15.3.11.161: C=new GetRequest(30) .1.3.6.1.4.1.2021.4.5.0
12:24:48.386451 IP 0.0.0.2.45812 > 10.15.3.11.1521: S 3199704795:3199704795(0) win 5760
Why it uses 0.0.0.2 as source address for monitoring?
LTM version 10.2.4 build 595.0
I have another LTM with same desing, but with version 10.2 - all works fine.
