Date: Sat, 21 Jul 2007 02:17:10 +0200 From: Stefan Esser <se@FreeBSD.org> To: Julian Elischer <julian@elischer.org> Cc: Xin LI <delphij@delphij.net>, hackers@freebsd.org, Stefan Esser <se@localhost.FreeBSD.org> Subject: Re: tunefs.8 oddity Message-ID: <46A15086.2060505@FreeBSD.org> In-Reply-To: <46A1410F.3080703@elischer.org> References: <20070720185101.F20123@mp2.macomnet.net> <46A0D313.8010904@delphij.net> <20070720192830.C20123@mp2.macomnet.net> <46A13A10.8070108@FreeBSD.org> <46A1410F.3080703@elischer.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Julian Elischer wrote: > Stefan Esser wrote: >> Maxim Konovalov wrote: >>> On Fri, 20 Jul 2007, 23:21+0800, Xin LI wrote: >>>> Any chance that we resolve the bug instead of documenting it? :-) >>>> >>> Personally, I have no energy/time for that. It was documented for >>> ages, it is still documented in other BSDs. >> >> It has long ago been converted to a mount option instead of a >> tunefs command in NetBSD. My FreeBSD systems are patched that >> way since shortly after soft-updates was committed (long before >> it became available in NetBSD, IIRC); I have not checked whether >> they used the (very simple) patches I had posted at that time. >> >> Controlling soft-updates during mount has many advantages (not >> only if you decide to enable it on a root file-system that had >> been created without it) and no disadvantages. >> > > As the person who added this originally on behalf of Kirk, > I think the time for this has probably come. > I think even Kirk has said this might now make sense but I'd check > with him first. I had asked him and he replied that the mount option was better. But when I discussed this in a FreeBSD list, there was strong opposition and it was claimed, that sysinstall took care of it in such a way, that the users would not miss the mount option. I had strong arguments in favour of the mount option, but since there were different opinions that were strongly voiced, I just gave up and maintained the changes locally to this day. The arguments (of both sides) can be found in the mail archives. I think this change should have been made long before 7.0 (it could have been in 5.0 without problems). But I guess that ZFS will attract many previous UFS2/soft-updates users (I have converted most of my systems to ZFS with quite satisfactory results and could now live without soft-updates) and that the relevance of soft-updates will shrink for that reason. So if it is not changed in 7.0, I won't keep my local patches, since there will be no UFS file system on my systems (I'm using a ZFS root partition and have only a UFS boot partition). Regards, STefan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?46A15086.2060505>