On 7/20/06, Cliff Wells <[EMAIL PROTECTED]> wrote:
Well, the other "well-known" bit of info is that ext3 gets much of its
"reliability" from syncing every 5 seconds. If you want to use XFS and
get that sort of data reliability, here's a bash script to add to
rc.local:
( while true; do sync; sleep 5; done )&
Well, you laugh, but my /etc/sysctl.conf contains:
vm.laptop_mode = 0
fs.xfs.xfssyncd_centisecs = 500
You can also mount XFS in sync mode if you are paranoid, but be warned
that it keeps your disks *very* busy.
Yeah. I would rather use ext3 with data=journal!
So I guess the real question is this: what qualifies as "FS
reliability".
Right. "Sucks" is imprecise in most circles.
But consider this...the entire value of a filesystem is the files it
contains. A filesystem that fixes itself by doing the equivalent of
"mkfs" on reboot from a crash will be both completely consistent, and
completely useless. By anyone's definition, it would "suck".
cross-linked files and bad inode counts). Also, having to fsck a large
disk array is going to be quite painful.
Yes, ext3 maintainers are well aware of this. Have you seen:
http://infohost.nmt.edu/~val/fs_workshop/
And the lwn article:
http://lwn.net/Articles/189547/
-Richard
--
[email protected] mailing list
