master->backup CARP issue

[Girish Venkatachalam]

On 09:50:51 Nov 14, Eugene wrote:
> Hello freebsd,
> 
> 
> We have two FreeBSD boxes, each one has 4 CARP interfaces in preempt
> mode, without arp balancing.
> 
> Backup without any reason preempts Master and after a while (an hour
> or two) it becomes Backup again.
> When Backup becomes Master on Master box
> net.inet.carp.suppress_preempt: 1, but all interfaces are up and
> seemed to work properly.
> 
> 
> Is there any ways to debug this issue?
> 
> 
> Additional info:
> 
> #uname -v
> FreeBSD 6.1-RELEASE-p7  
> 
> Master:
> # grep carp /var/log/messages
> Nov 14 01:50:59 atlas kernel: carp3: MASTER -> BACKUP (more frequent advertisement received)
> Nov 14 01:52:30 atlas kernel: carp3: MASTER -> BACKUP (more frequent advertisement received)
> Nov 14 02:43:55 atlas kernel: carp3: MASTER -> BACKUP (more frequent advertisement received)
> Nov 14 02:59:12 atlas kernel: carp3: MASTER -> BACKUP (more frequent advertisement received)
> Nov 14 03:26:13 atlas kernel: carp3: MASTER -> BACKUP (more frequent advertisement received)
> Nov 14 03:26:32 atlas kernel: carp3: MASTER -> BACKUP (more frequent advertisement received)
> Nov 14 04:58:39 atlas kernel: carp3: MASTER -> BACKUP (more frequent advertisement received)
> Nov 14 05:53:27 atlas kernel: carp3: MASTER -> BACKUP (more frequent advertisement received)
> Nov 14 06:01:12 atlas kernel: carp3: MASTER -> BACKUP (more frequent advertisement received)
> Nov 14 06:03:36 atlas kernel: carp2: BACKUP -> MASTER (preempting a slower master)
> Nov 14 06:03:36 atlas kernel: carp0: BACKUP -> MASTER (preempting a slower master)
> Nov 14 06:03:36 atlas kernel: carp1: BACKUP -> MASTER (preempting a slower master)
> Nov 14 06:03:38 atlas kernel: carp3: BACKUP -> MASTER (preempting a slower master)
> 
> Master & Backup (in normal state)
> #sysctl -a | grep "net.inet.carp."
> net.inet.carp.allow: 1
> net.inet.carp.preempt: 1
> net.inet.carp.log: 2
> net.inet.carp.arpbalance: 0
> net.inet.carp.suppress_preempt: 0
> 
> all NICs - 3Com 3c905C-TX
> 
> Master:
> #ifconfig | grep -A2 carp
> carp0: flags=49<UP,LOOPBACK,RUNNING> mtu 1500
>         inet 10.1.1.3 netmask 0xffffff00
>         carp: MASTER vhid 1 advbase 1 advskew 10
> carp1: flags=49<UP,LOOPBACK,RUNNING> mtu 1500
>         inet 10.1.2.3 netmask 0xffffff00
>         carp: MASTER vhid 2 advbase 1 advskew 10
> carp2: flags=49<UP,LOOPBACK,RUNNING> mtu 1500
>         inet 10.1.3.3 netmask 0xffffff00
>         carp: MASTER vhid 3 advbase 1 advskew 10
> carp3: flags=49<UP,LOOPBACK,RUNNING> mtu 1500
>         inet 10.1.4.3 netmask 0xfffffff8
>         carp: MASTER vhid 4 advbase 1 advskew 10
> 
> Backup
> #ifconfig |grep -A2 carp
> carp0: flags=49<UP,LOOPBACK,RUNNING> mtu 1500
>         inet 10.1.1.3 netmask 0xffffff00
>         carp: BACKUP vhid 1 advbase 1 advskew 20
> carp1: flags=49<UP,LOOPBACK,RUNNING> mtu 1500
>         inet 1.1.2.3 netmask 0xffffff00
>         carp: BACKUP vhid 2 advbase 1 advskew 20
> carp2: flags=49<UP,LOOPBACK,RUNNING> mtu 1500
>         inet 10.1.3.3 netmask 0xffffff00
>         carp: BACKUP vhid 3 advbase 1 advskew 20
> carp3: flags=49<UP,LOOPBACK,RUNNING> mtu 1500
>         inet 10.1.4.3 netmask 0xfffffff8
>         carp: BACKUP vhid 4 advbase 1 advskew 20
> 
> Master:
> 
> #netstat -ni |grep carp
> Name    Mtu Network       Address              Ipkts Ierrs    Opkts Oerrs  Coll
> carp0  1500 <Link#12>                         311894     1  5926852     0     0
> carp0  1500 10.1.1        10.1.1.3          42293866     -   263238     -     -
> carp1  1500 <Link#13>                          48653     1  5663615     0     0
> carp1  1500 10.1.2        10.1.2.3         153785957     -        0     -     -
> carp2  1500 <Link#14>                          48661     1  5663615   107     0
> carp2  1500 10.1.3/24     10.1.3.3            476782     -        0     -     -
> carp3  1500 <Link#15>                          48974     1  5663634  1155     0
> carp3  1500 10.1.4.3/     10.1.4.3           6885622     -        0     -     -
>         
> 
> Backup:
> 
> #netstat -ni |grep carp
> Name    Mtu Network       Address              Ipkts Ierrs    Opkts Oerrs  Coll
> carp0  1500 <Link#12>                          35221     0     6437     0     0
> carp0  1500 10.1.1        10.1.1.3              8699     -        0     -     -
> carp1  1500 <Link#13>                          35220     0     6438     0     0
> carp1  1500 10.1.2        10.1.2.3             86000     -        0     -     -
> carp2  1500 <Link#14>                          35221     0     6437     0     0
> carp2  1500 10.1.3/24     10.1.3.3                 0     -        0     -     -
> carp3  1500 <Link#15>                          35190     0     6443    23     0
> carp3  1500 10.1.4.3/     10.1.4.3               958     -        0     -     -
>         
> 
> Thanks in advance!

I could be talking nonsense but still...

Have you tried adjusting the "advskew" and "advbase" parameters?

Another thing. Are the NICs reliable?

Have you tried with a different NIC?

Looks like you don't have a faulty NIC here. So it could be the other
possibility.

Thanks.

regards,
Girish