Date: Tue, 22 Apr 2008 14:36:26 -0400 From: "Zaphod Beeblebrox" <zbeeble@gmail.com> To: "Andrew Snow" <andrew@modulus.org> Cc: freebsd-stable@freebsd.org, Pete French <petefrench@ticketswitch.com> Subject: Re: Dreadful gmirror performance, though each half works fine Message-ID: <5f67a8c40804221136s2c1893c0tdd00c627ab813c59@mail.gmail.com> In-Reply-To: <480D7F58.1080203@modulus.org> References: <E1JnuQf-000LaE-5n@dilbert.ticketswitch.com> <480D7F58.1080203@modulus.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Apr 22, 2008 at 2:02 AM, Andrew Snow <andrew@modulus.org> wrote: > Pete French wrote: > > > I did some benchmarking, and "load" gives me a bit better performance > > than > > "round-robin" so I've elected to use that. Haven't tried "prefer" as > > syncing all the drives backwards and forwards to get the preferences set > > seems a bit too much like hard work! > > > > I use this patch for sbin/geom/class/mirror/geom_mirror.c > > Change: > md.md_priority = i - 1; > To: > md.md_priority = i - 1 + 100; I hate to ask for the "right" solution, but shouldn't we be patching the gmirror userland to accept a priority argument to label and make the kernel part listen to that? This patch does make sense --- but it doesn't go far enough. Also, it seems sensible that you should be able to modify the priority values of a running disk.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5f67a8c40804221136s2c1893c0tdd00c627ab813c59>