On Tue, 08 Aug 2006 14:53:32 +1200 Volker Kuhlmann <[EMAIL PROTECTED]> wrote:
[snip] > reiserfs on at least / and /home on all systems for uhm 5-7 years now > and the recovery rate is 100%. The bootup-fsck fixes it for me (that's > the reason one uses journalling filesystems). Ext3 back then wasn't > really available (yes ext3 was a latecomer) and ext2 is unsuitable for > anything much larger than a gigabyte if you don't want to die of old age > during fsck, there wasn't that much choice actually. > > I remember fixing one bad superblock with fsck but don't recall which > filesystem, I was using ext2 as well as reiser for some time, nor do I > remember whether the cause was a disk failure. > > > Just one 'bad superblock' experience (under Gentoo at the time) was > > enough to put me off for life. :) [snip] Must admit that the only two times that a filesystem has become so completely unravelled as to be unrecoverable, the score is 1 to Resier and 1 to SGI's XFS. Just one query though. Given that most IO intensive operations make heavy use of the Gigabytes of spare memory that we all now have, what difference would that make to the security of things like copying loads of data??? The file system *shouldn't* make a difference, but I was wondering what was marked read/written where and when. Anyone got any insights? Steve
