original master comes up and takes back VIP

[Gilad Abada]

Hi Everyone

I have 2 Dell 2950 servers running ubuntu 10.04 LTS. I installed
ucarp with apt-get install ucarp.

The issue I am having is that one of the machines (whether its set
for master or not) always takes control of the VIP when it is up.

I would like for the system with control of the VIP to keep it until it
goes down and the other system will pick it up.

Is there a setting for this? I have this working on other systems
but can't seem to get it to work on these.

In the log of the machine that steals the VIP back I get this:
Jan 8 18:17:53 lennon kernel: [ 409.557687] bnx2: eth1 NIC Copper Link is Up, 1000 Mbps full duplex, receive & transmit flow control ON
Jan 8 18:18:23 lennon ucarp[1495]: [WARNING] Non-preferred master advertising: reasserting control of VIP with another gratuitous arp
Jan 8 18:18:23 lennon kernel: [ 439.888877] IN=eth0 OUT= MAC=01:00:5e:00:00:12:00:XX:XX:00:00:XX:XX:XX SRC=XXX.XXX.XXX.XXX DST=224.0.0.18 LEN=56 TOS=0x10 PREC=0x00 TTL=255 ID=29489 DF PROTO=112
Jan 8 18:18:44 lennon ucarp[1495]: [ERROR] exiting: pfds[0].revents = 8
Jan 8 18:18:44 lennon ucarp[1495]: [WARNING] Spawning [/usr/share/ucarp/vip-down eth0 XXX.XXX.XXX.XXX]
Jan 8 18:18:44 lennon kernel: [ 460.850114] bnx2 0000:05:00.0: irq 97 for MSI/MSI-X
Jan 8 18:18:45 lennon kernel: [ 461.021131] bnx2: eth0: using MSI
Jan 8 18:18:45 lennon kernel: [ 461.021808] ADDRCONF(NETDEV_UP): eth0: link is not ready
Jan 8 18:18:45 lennon init: ssh main process (1564) terminated with status 255
Jan 8 18:18:45 lennon ucarp[1730]: [INFO] Local advertised ethernet address is [00:1e:XX:XX:XX:XX]
Jan 8 18:18:45 lennon kernel: [ 461.081685] bnx2 0000:09:00.0: irq 98 for MSI/MSI-X
Jan 8 18:18:45 lennon ucarp[1730]: [WARNING] Switching to state: BACKUP
Jan 8 18:18:45 lennon ucarp[1730]: [WARNING] Spawning [/usr/share/ucarp/vip-down eth0 XXX.XXX.XXX.XXX]
Jan 8 18:18:45 lennon kernel: [ 461.280019] bnx2: eth1: using MSI
Jan 8 18:18:45 lennon kernel: [ 461.280793] ADDRCONF(NETDEV_UP): eth1: link is not ready
Jan 8 18:18:45 lennon init: ssh main process (1726) terminated with status 255
Jan 8 18:18:47 lennon kernel: [ 463.656580] bnx2: eth1 NIC Copper Link is Up, 1000 Mbps full duplex, receive & transmit flow control ON
Jan 8 18:18:47 lennon kernel: [ 463.657205] ADDRCONF(NETDEV_CHANGE): eth1: link becomes ready
Jan 8 18:18:47 lennon kernel: [ 463.790754] bnx2: eth0 NIC Copper Link is Up, 1000 Mbps full duplex
Jan 8 18:18:47 lennon kernel: [ 463.791368] ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
Jan 8 18:18:58 lennon kernel: [ 474.290041] eth0: no IPv6 routers present
Jan 8 18:18:58 lennon kernel: [ 474.600026] eth1: no IPv6 routers present
Jan 8 18:18:59 lennon ucarp[1730]: [WARNING] Switching to state: MASTER
Jan 8 18:18:59 lennon ucarp[1730]: [WARNING] Spawning [/usr/share/ucarp/vip-up eth0 XXX.XXX.XXX.XXX]
Jan 8 18:18:59 lennon init: ssh main process (1764) terminated with status 255
Jan 8 18:19:17 lennon ucarp[1730]: [WARNING] Non-preferred master advertising: reasserting control of VIP with another gratuitous arp

We also need:

1) The release of Ubuntu you are using, via 'lsb_release -rd' or System -> About Ubuntu
Description: Ubuntu 10.04.3 LTS
Release: 10.04

2) The version of the package you are using, via 'apt-cache policy pkgname' or by checking in Software Center
ucarp:
  Installed: 1.5.1-1ubuntu0.1
  Candidate: 1.5.1-1ubuntu0.1
  Version table:
 *** 1.5.1-1ubuntu0.1 0
        500 http://us.archive.ubuntu.com/ubuntu/ lucid-updates/universe Packages
        100 /var/lib/dpkg/status
     1.5.1-1 0
        500 http://us.archive.ubuntu.com/ubuntu/ lucid/universe Packages

3) What you expected to happen
I expect the new master to stay as master until it fails. Old master should come back up and slave.

4) What happened instead
The Old master comes up and takes the IP back

Thanks!
Gill