I'm working on CARP with IP balancing on 4.6R. With trunk(4) failover
setting, it
doesn't work.
# cat hostname.em0
up
# cat hostname.em1
up
# cat hostname.trunk1
trunkport em0 trunkport em1
trunkproto failover
up
# cat hostname.carp0
carpdev trunk1
carpnodes 72:0,172:100 balancing ip-stealth 158.205.129.72/28
up
!route add default `cat /etc/mygate`
!arp -d 158.205.129.67
!arp -s 158.205.129.67 00:00:0c:07:ac:00 permanent
# ifconfig carp0
carp0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
lladdr 00:00:5e:00:01:48
priority: 0
carp: carpdev trunk1 advbase 1 balancing ip-stealth
state MASTER vhid 72 advskew 0
state BACKUP vhid 172 advskew 100
groups: carp egress
inet6 fe80::200:5eff:fe00:148%carp0 prefixlen 64 scopeid 0xb
inet 158.205.129.72 netmask 0xfffffff0 broadcast 158.205.129.79
On the other node, configuration is almost identical, other than vhid and
advskew pairs. So far, so good. But when em0 is down, carp0 incorrectly
detects link state change and vhid 172 becomes MASTER (wtf?).
# ifconfig em0 down
# ifconfig carp0
carp0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
lladdr 00:00:5e:00:01:48
priority: 0
carp: carpdev trunk1 advbase 1 balancing ip-stealth
state MASTER vhid 72 advskew 0
state MASTER vhid 172 advskew 100
groups: carp egress
inet6 fe80::200:5eff:fe00:148%carp0 prefixlen 64 scopeid 0xb
inet 158.205.129.72 netmask 0xfffffff0 broadcast 158.205.129.79
The other node is still BACKUP (vhid 72) and MASTER (vhid 172). Now vhid 172 is
MASTER-MASTER state.
Am I mssing something? Maybe fixed in -current?
--
Tomoyuki Sakurai
