[Grml] Write barriers for journaling filesystems
Mark
27e3kk302 at sneakemail.com
Sun Jan 21 04:32:26 CET 2007
Testing exercises with XFS, ext3, and power failure. I found advice to
use write barriers. It appears that any journaling file system
(XFS/Reiser/ext3) will benefit w.r.t. power failures. Almost all hard
drives, IDE, SATA, USB, SCSI have a ~multimegabyte write cache nowadays.
"OK, enough theory. Do this: mount all of your filesystems with I/O
barrier turned on. Something like this (example line from my
/etc/fstab): /dev/sda5 / ext3 defaults,barrier=1 1 1"
http://kerneltrap.org/node/6173#comment-193357
"Fortunately, you can have your cake and eat it too. The trick is
to implement IO barriers using the CACHE FLUSH and/or FUA commands.
That way you can have the performance and MTBF benefits of
write-back caching while still having a safe fsync() (safe as in
doesn't return before data is on the platters)."
http://lwn.net/Articles/181393/
It seems like maybe a good default setting for journaled filesystems,
especially on USB (which users often yank out...).
M
P.S. Update on ZFS. Apple will include it in their next OS release
(mika: not as default/boot, still HFS+ for that). Also, Sun may release
ZFS under GPLv3 "eventually."
http://apple.slashdot.org/apple/06/12/18/1155226.shtml
http://www.itweek.co.uk/vnunet/news/2172823/sun-kills-gpl3-rumours
More information about the Grml
mailing list