Date: Wed, 17 Jan 2007 15:07:49 -0800 (PST) From: Tom Samplonius <tom@samplonius.org> To: msaad@datapipe.com Cc: freebsd-stable@freebsd.org Subject: Re: Failover-HA-Setup Message-ID: <8537645.61169075269417.JavaMail.root@ly.sdf.com> In-Reply-To: <45AEA2F6.9020204@datapipe.com>
next in thread | previous in thread | raw e-mail | index | archive | help
----- Mark Saad <msaad@datapipe.com> wrote: > Ivan Voras wrote: > > Dmitriy Kirhlarov wrote: > >> On Wed, Jan 17, 2007 at 03:20:32PM +0100, Richard Verwayen wrote: > >> > >>> Oh, should I have mentioned that it is a shared SCSI-Storage in > between? > >> You can use gmirror(8) with ggated(8). > > > > I think his problem is that you can't have two machines RW mounting > the > > same device (doesn't matter if it's FC, ggate or other...). So when > one > > machine fails, the other must mount the file system and THEN start > > daemons... > > > > > > > This is kind of a side topic but does anyone know of a on disk > clustered > file system for FreeBSD; like OCFS2 or GFS . Both which would allow > you > to have more then one box attached read / write to the same scsi or fc > > storage array ? I don't believe there are any. Additionally, I don't believe any of the clustered file systems would be usable for a database backend. Because the application does not know that another instance is also updating, so you would have to start only one instance at a time. So the locking overhead of a clustered filesystem isn't worth it for databases. > -- > Mark Saad > msaad@datapipe.com
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?8537645.61169075269417.JavaMail.root>