sf...@users.sourceforge.net:
> > Do you consider this a bug in aufs, though?
>
> No.
> I think it is a problem of your shutdown script.
> Generally any shutdown script executes
> - kill all processes
> - remount / readonly
>
> For the system whose root is aufs, the similar scirpt but more work is
> necessary. In your case,
> - remount ext3 readwrite
> - kill all processes
> - remount / readonly
> I guess.
I got another question in my mind.
At the end of your rsync script,
mount -o remount,ro ${LOWER_BRANCH}
Why didn't this remount return an error?
By rsync, a file might be renamed. So the inode became
- its link count is zero.
- its reference count is positive, it is still alive.
- its dtime is not set yet. it is set when the reference count reaches
zero (in ext3).
For such inode, fsck _always_ reports "Deleted inode XXXX has zero
dtime" definetly.
As long as a process keeps opening a file, aufs has to keep the
corresponding file on a branch opened too. It means the inode reference
count is kept positive, and its dtime keeps zero until the file is
closed.
Why does ext3 allow to be readonly in the status which fsck thinks
error?
J. R. Okajima
------------------------------------------------------------------------------
Increase Visibility of Your 3D Game App & Earn a Chance To Win $500!
Tap into the largest installed PC base & get more eyes on your game by
optimizing for Intel(R) Graphics Technology. Get started today with the
Intel(R) Software Partner Program. Five $500 cash prizes are up for grabs.
http://p.sf.net/sfu/intelisp-dev2dev
