Date: Sat, 4 Sep 2004 23:24:14 -0300 (ADT) From: "Marc G. Fournier" <scrappy@hub.org> To: Dan Nelson <dnelson@allantgroup.com> Cc: freebsd-questions@freebsd.org Subject: Re: File system replication between servers Message-ID: <20040904232130.O812@ganymede.hub.org> In-Reply-To: <20040905015820.GA44261@dan.emsphone.com> References: <20040904222712.F812@ganymede.hub.org> <20040905015820.GA44261@dan.emsphone.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, 4 Sep 2004, Dan Nelson wrote:
> In the last episode (Sep 04), Marc G. Fournier said:
>> Does anyone know of any software that will run on FreeBSD that would
>> allow you to keep two servers in sync? All writes to /dir1 on
>> server1 would go to /dir1 on server2, and all writes to /dir2 on
>> server2 would go to /dir2 on server1?
>>
>> I've thought about rsync, but am trying to get it as close to real
>> time as I can, and on large file systems, rsync would just get too
>> far behind, too quickly ...
>>
>> I've started to look at Coda6, but so far what I'm reading through
>> the docs makes it sound like it may not be suitable ...
>>
>> What I'm looking for is a hot failover solution using FreeBSD ... if
>> server1 goes down, server2 can start up and continue working ... when
>> server1 comes back up again, changes to server2 could be re-sync'd
>> back to server1, and server1 could once more resume its duties ...
>
> How about a shared SCSI drive, filesystems only mounted on the master.
> When the master fails, the slave fscks the filesystems, mounts them,
> and becomes the master. Tried and true. You could even use background
> fsck and get immediate failover.
need to do this above the hardware layer ... am looking at 50% of services
running on each of 2 servers, with failover to 100% running on one of the
servers if the other goes down ...
reading through coda's docs, I think I can do it using that, just curious
if there is a simpler setup that could be used ...
> You could even do it without shared storage if you use geom_gate and
> geom_{mirror,vinum,ccd} to keep two identical disks on each machine in
> sync. When the master crashes and comes back up as a slave, the mirror
> code will update its disks.
what is geom_gate? checked 'man geom_gate' on my 5.x machine, and nothing
shows up ...
----
Marc G. Fournier Hub.Org Networking Services (http://www.hub.org)
Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040904232130.O812>