http://oss.sgi.com/bugzilla/show_bug.cgi?id=721
------- Additional Comments From tes@xxxxxxx 2006-10-02 21:40 CST -------
Torsten,
You confused me with the bit about being fine in xfsprogs 2.7.11 and broken now.
What you were referring to was:
xfsprogs-2.7.13 (08 February 2006)
- Convert fsck into a shell script and make it provide a
hint to run repair/check (in non-auto fsck modes).
Thanks to Marian Jancar <mjancar@xxxxxxx> for this.
I take it that you were thinking that prior to 2.7.13 we were calling xfs_check
(??)
since we didn't print out this message - when, in fact, AFAIK we never did.
I see your point, but like Russell said, it is too slow.
(Certainly slow for really large filesystems)
And that was one of the motivations of journalled filesystems AFAIK (that
fsck on boot wasn't paramount).
I guess there are different kinds of corruption.
The typical ones of missing metadata due to unclean unmounts (e.g. power
failure)
are handled by the metadata log and its replay.
The (hopefully) less typical ones of software bugs and hardware failures etc..
will come apparent with log messages, filesystem shutdown, syscall failures and
if you run
check/repair.
--Tim
--
Configure bugmail: http://oss.sgi.com/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
|