Sorry, perhaps we are missing any point.
I was testing exporting XFS, and now I've repeated
the test with ext2. (using my /boot ;) )
I can't imagine what is wrong:
[EMAIL PROTECTED]:/# exportfs -o no_root_squash localhost:/boot
[EMAIL PROTECTED]:/# mount localhost:/boot /mnt
[EMAIL PROTECTED]:/# mount |grep mnt
localhost:/boot on /mnt type nfs (rw,addr=127.0.0.1)
[EMAIL PROTECTED]:/# cd /mnt
[EMAIL PROTECTED]:/mnt# mkdir A; touch A/B; ln A/B B; tail -f A/B &
[1] 9267
[EMAIL PROTECTED]:/mnt# rm A/B; rmdir A
rmdir: A: Directory not empty
[EMAIL PROTECTED]:/mnt# echo AAAA >> B
[EMAIL PROTECTED]:/mnt# AAAA
With aufs the exactly same test has a diferent behaviour,
But in your tests, the "rm A/B" is truncating the file? Even
existing the hard link?
Thanks,
Bona
p.s.: The problem is that I would like to export via NFS an
AUFS root including the /tmp. Actually I export a readonly
root, then mount the new root using aufs and chroot to it.
But with a /tmp being AUFS exported via NFS, mutt and
some other apps didn't work correctly.
The test was extracted "stracing" mutt lock:
open("/tmp/.muttIOAHhM/mutt-bowmore-1413-22816-0",
O_WRONLY|O_CREAT|O_EXCL|O_LARGEFILE|O_NOFOLLOW, 0600) = 4
link("/tmp/.muttIOAHhM/mutt-bowmore-1413-22816-0",
"/tmp/mutt-bowmore-1413-22816-0") = 0
stat64("/tmp/.muttIOAHhM/mutt-bowmore-1413-22816-0", {st_mode=S_IFREG|0600,
st_size=0, ...}) = 0
stat64("/tmp/mutt-bowmore-1413-22816-0", {st_mode=S_IFREG|0600, st_size=0,
...}) = 0
unlink("/tmp/.muttIOAHhM/mutt-bowmore-1413-22816-0") = 0
unlink("/tmp/.muttIOAHhM/mutt-bowmore-1413-22816-0") = -1 ENOENT (No such file
or directory)
rmdir("/tmp/.muttIOAHhM") = -1 ENOTEMPTY (Directory not empty)
lstat64("/tmp/mutt-bowmore-1413-22816-0", {st_mode=S_IFREG|0600, st_size=0,
...}) = 0
fstat64(4, {st_mode=S_IFREG|0600, st_size=0, ...}) = 0
fcntl64(4, F_GETFL) = 0x28001 (flags
O_WRONLY|O_LARGEFILE|O_NOFOLLOW)
fstat64(4, {st_mode=S_IFREG|0600, st_size=0, ...}) = 0
mmap2(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) =
0xb741c000
_llseek(4, 0, [0], SEEK_CUR) = 0
_llseek(3, 0, [62540613], SEEK_CUR) = 0
On Tue, Oct 09, 2007 at 11:02:59AM +0900, [EMAIL PROTECTED] wrote:
>
> "Luis Carlos Erpen de Bona":
> > > What will happen if you export ext2 instead of aufs, execute tail and
> > > rm? Is the behaviour is different?
> > Yes. Even in NFS it works correctly. If the file still
> > existing (with a hard link, the inode count is not 0 when
> > you delete A/B) the file descriptor should stay working.
>
> What is the behaivour you expected in 'Even in NFS it works correctly'?
> I've tested your command sequence (of course, including 'ln') on nfs
> which is exported ext2. And the behaviour is equal to the exported aufs.
>
>
> Junjiro Okajima
-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
