Date: Sat, 5 Feb 2005 09:40:59 -0800 From: Steve Kargl <sgk@troutmask.apl.washington.edu> To: Robert Watson <rwatson@FreeBSD.org> Cc: freebsd-chat@FreeBSD.org Subject: Re: the value of a journal filesystem? Message-ID: <20050205174059.GA91377@troutmask.apl.washington.edu> In-Reply-To: <Pine.NEB.3.96L.1050205172810.55669D-100000@fledge.watson.org> References: <20050204192511.GA84359@troutmask.apl.washington.edu> <Pine.NEB.3.96L.1050205172810.55669D-100000@fledge.watson.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, Feb 05, 2005 at 05:30:26PM +0000, Robert Watson wrote: > > On Fri, 4 Feb 2005, Steve Kargl wrote: > > > I'm not sure if this is a ext2fs, ext3fs, or reiserfs, but the 2nd > > paragraph is somewhat ominious. The notice does statement whether the > > damaged filesystems are on other disks or on disks in other machines > > (ie. nfs mounted). > > Journalling, as with Soft Updates, relies on generally correct operation > of the media (i.e., changes are written or not, etc), and is intended to > protect only against "fail stop" failure modes. Handling media failure is > generally a task for RAID arrays, which are intended to mask corruption by > coercing corruption to "fail stop" on the media. So the interesting > question here would be: did their RAID not protect them? Or did they not > have RAID? > >From what I've read, redhat replaced a dead disk in a raid array with a new disk and started a recovery phase. During recovery, the filesystems were corrupted. I have not been able to find any info on what hardware controller redhat uses (or used :). gcc.gnu.org has been down for 72+ hours, which seems like a very long time for such an important site. -- Steve
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050205174059.GA91377>