I created a virtual IP that will be used for failover of a mysql DB with the following command:
pcs resource create TESTDB_vip IPaddr2 ip=10.1.16.222 cidr_netmask=24 nic=vlan16 iflabel=TESTDB noprefixroute=true flush_routes=false op monitor interval=30s
The resource starts ok on one node, but on failover to the 2nd node, I get error:
ERROR: More than 1 routes match 10.1.16.222/24. Unable to decide which route to use.
WARNING: [findif] failed
On the node where it fails, besides the primary IP there are also 2 other secondary IPs that are managed by NetworkManager, where the /etc/NetworkManager/system-connections/vlan16.nmconnection includes:
***\[ipv4\]
address1=10.1.16.232/24
address2=10.1.16.145/24
address3=10.1.16.146/24
And “ip route” includes entries for each of those IP addresses on bridge vlan16 and 'ip addr” shows address2 and address3 as “secondary”
21: vlan16 inet 10.1.16.232/24 brd 10.1.16.255 scope global noprefixroute vlan16\ valid_lft forever preferred_lft forever
21: vlan16 inet 10.1.16.145/24 brd 10.1.16.255 scope global secondary noprefixroute vlan16\ valid_lft forever preferred_lft forever
21: vlan16 inet 10.1.16.146/24 brd 10.1.16.255 scope global secondary noprefixroute vlan16\ valid_lft forever preferred_lft forever
The findIF.sh is confused because of the multiple route entries for interface “vlan16”
Beside modifying the findif.sh, are there any other good work-arounds?
Am I missing anything in my pacemaker configuration?
Thanks in advance