HAST, ZFS and CARP [Re: freebsd-fs Digest, Vol 357, Issue 3]

[Davide D'Amico]

> Message: 3
> Date: Tue, 20 Apr 2010 07:39:43 -0700
> From: Freddie Cash <fjwcash at gmail.com>
> Subject: Re: HAST and ZFS
> To: freebsd-fs at freebsd.org
> Message-ID:
>        <n2nb269bc571004200739pd11f7121y5eecb24e906e17cd at mail.gmail.com>
> Content-Type: text/plain; charset=UTF-8
>
> On Tue, Apr 20, 2010 at 7:24 AM, Davide D'Amico <davide.damico at gmail.com>wrote:
>
>> I'm trying to setup an active/passive cluster.
>> Some questions:
>> - why the use of net/ucarp rather than carp? Is it for its up/down scripts?
>> - I would like to use ZFS on top on HAST, so I think that in
>> wiki.freebsd.org/HAST
>> I should change UFS (in sh scripts) but how to replace newfs command?
>> With zpool create disk da0?
>>
>
> See the mailing list archives.  I posted a devd.conf and script for using
> CARP instead of uCARP, for using HAST devices to create a ZFS pool.  Works
> quite nicely, although there are no checks/fixes to prevent split-brain if
> CARP flip-flops between the two hosts.

Hi,
I've read the mailling list archives (lists.freebsd.org give me an
internal server error),
and I've successfully created a failover cluster using hast, zfs and carp.
I tried to shutdown the primary node, and than I see on he secondary node:

Apr 21 15:27:45 hastb hastd: Connection from tcp4://10.8.0.3:8457 to
tcp4://10.8.0.3:55949.
Apr 21 15:27:45 hastb hastd: [da0s2] (primary) We act as primary for
the resource and not as secondary as requested by
tcp4://10.8.0.3:55949.
Apr 21 15:27:45 hastb hastd: [da0s2] (primary) Remote node acts as
primary for the resource and not as secondary.

How could I resolve these issues?

Thanks,
d.