Hi everybody,
I am a happy user of aufs for our Gibraltar firewall distribution. However,
there is one issue that I have not managed to solve so far. aufs code has been
taken from CVS head as of today (2009-04-17) with kernel 2.6.28.9 with only a
few patches (including PaX).
[~]# modinfo aufs
filename: /lib/modules/2.6.28.9/kernel/fs/aufs/aufs.ko
version: 20090126
description: aufs -- Another unionfs
author: Junjiro Okajima
license: GPL
srcversion: E3370452FEAC89EDB52A420
depends: exportfs
vermagic: 2.6.28.9 SMP preempt mod_unload 586
parm: nwkq:the number of workqueue thread, aufsd (short)
parm: brs:use <sysfs>/fs/aufs/si_*/brN (int)
By default, /etc and /var are mounted with aufs:
/dev/loop0 on /system/root type squashfs (ro,noatime)
tmpfs on /system/ramdisks/etc type tmpfs (rw,size=8192k,mode=755)
aufs on /etc type aufs
(rw,si=158ccebd,xino=/system/ramdisks/etc/.aufs.xino,br:/system/ramdisks/etc=rw:/system/root/etc=rr)
tmpfs on /system/ramdisks/var type tmpfs (rw,mand,size=32768k,mode=755)
aufs on /var type aufs
(rw,mand,si=145a95fd,xino=/system/ramdisks/var/.aufs.xino,br:/system/ramdisks/var=rw:/system/root/var=rr)
tmpfs on /var/tmp type tmpfs (rw,mand,size=32768k,mode=755)
with tmpdisk as writeable and squashfs (LZMA compressed) as the readonly
branch.
The problem is now a specific directory under /etc:
[~]# touch /etc/gibraltar/test
[ 1574.915789] aufs au_lkup_neg:393:touch[6175]: I/O Error, b0 gibraltar
should be negative.
touch: cannot touch `/etc/gibraltar/test': Input/output error
while all other directories under /etc are properly writable:
[~]# touch /etc/test
[~]# touch /etc/test2
[~]# touch /etc/network/test
[~]# touch /etc/init.d/test4
[~]# echo test > /etc/event.d/foo
I don't see anything special with that directory:
[~]# ls -ld /etc/gibraltar/
drwxr-xr-x 3 root root 67 Mar 24 19:15 /etc/gibraltar//
[~]# ls -l /etc/gibraltar/
total 4
-rw-r--r-- 1 root root 2164 Mar 24 19:15 config
drwxr-xr-x 2 root root 3 Sep 16 2008 patches/
-rw-r--r-- 1 root root 328 Sep 16 2008 update-excludes
-rw-r--r-- 1 root root 13 Apr 17 11:54 version
which is a direct view of the underlying readonly branch:
[~]# ls -ld /system/root/etc/gibraltar/
drwxr-xr-x 3 root root 67 Mar 24 19:15 /system/root/etc/gibraltar//
[~]# ls -l /system/root/etc/gibraltar/
total 4
-rw-r--r-- 1 root root 2164 Mar 24 19:15 config
drwxr-xr-x 2 root root 3 Sep 16 2008 patches/
-rw-r--r-- 1 root root 328 Sep 16 2008 update-excludes
-rw-r--r-- 1 root root 13 Apr 17 11:54 version
Reading from it works:
[~]# cat /etc/gibraltar/version
2.99alpha065
Interestingly, when a script tries to write to this directory with
echo `date +%s` > /etc/gibraltar/config_lastsaved
echo `date` >> /etc/gibraltar/config_lastsaved
I get the error
[ 1975.325099] aufs au_lkup_neg:393:save-config[7847]: I/O Error, b0 gibraltar
should be negative.
/sbin/save-config: line 609: /etc/gibraltar/config_lastsaved: Input/output error
but the file appears in the read-write branch:
[~]# ls -la /system/ramdisks/etc/gibraltar/
total 4
drwxr-xr-x 2 root root 60 Apr 17 09:37 ./
drwxr-xr-x 15 root root 480 Apr 17 11:56 ../
-rw------- 1 root root 41 Apr 17 09:37 config_lastsaved
[~]# cat /system/ramdisks/etc/gibraltar/config_lastsaved
1239953831
Fri Apr 17 09:37:11 CEST 2009
but not in the aufs tree:
[~]# ls -l /etc/gibraltar/
total 4
-rw-r--r-- 1 root root 2164 Mar 24 19:15 config
drwxr-xr-x 2 root root 3 Sep 16 2008 patches/
-rw-r--r-- 1 root root 328 Sep 16 2008 update-excludes
-rw-r--r-- 1 root root 13 Apr 17 11:54 version
So, I am stumped and currently don't even have a clue on where to start
looking. Is this a bug in the kernel module or something stupid on my side?
The only difference between this directory and others that I can think of is
that it is the first that the bootup script tries to write to immediately after
having mounted the /etc tree with aufs. The state immediately after mounting
/etc is:
rootfs on / type rootfs (rw)
none on /sys type [ 88.348768] aufs au_lkup_neg:393:prepareroot[1119]: I/O
Error, b0 gibraltar should be negative.
sysfs (rw,nosuid,nodev,noexec)
none on /proc type proc (rw,nosuid,nodev,noexec)
udev on /dev type tmpfs (rw,size=10240k,mode=755)
/dev/hda1 on /live/image type vfat
(ro,noatime,fmask=0022,dmask=0022,codepage=cp437,iocharset=utf8)
/dev/loop0 on /gibraltar.squashfs type squashfs (ro,noatime)
tmpfs on /live/cow type tmpfs (rw,noatime,mode=755)
/dev/loop0 on / type squashfs (ro,noatime)
tmpfs on /live type tmpfs (rw)
none on /proc type proc (rw,nosuid,nodev,noexec)
none on /sys type sysfs (rw,nosuid,nodev,noexec)
udev on /dev type tmpfs (rw,size=10240k,mode=755)
none on /var/tmp type tmpfs (rw,size=16384k)
/dev/sda1 on /mnt/configmedia type vfat
(ro,fmask=0022,dmask=0022,codepage=cp437,iocharset=utf8)
/dev/loop0 on /system/root type squashfs (ro,noatime)
tmpfs on /system/ramdisks type tmpfs (rw,mand,size=256k,mode=755)
tmpfs on /system/ramdisks/etc type tmpfs (rw,size=8192k,mode=755)
aufs on /etc type aufs
(rw,si=158ccebd,xino=/system/ramdisks/etc/.aufs.xino,br:/system/ramdisks/etc=rw:/system/root/etc=rr)
Which then causes the same error while trying to write to
/etc/gibraltar/config_source:
/sbin/prepareroot: line 77: /etc/gibraltar/config_source: Input/output error
which indeed doesn't appear under /etc:
total 4
-rw-r--r-- 1 root root 2164 Mar 24 19:15 config
drwxr-xr-x 2 root root 3 Sep 16 2008 patches
-rw-r--r-- 1 root root 328 Sep 16 2008 update-excludes
-rw-r--r-- 1 root root 13 Apr 17 11:54 version
I would highly appreciate any pointers on how this can be fixed.
best regards,
Rene
--
-------------------------------------------------
Gibraltar firewall http://www.gibraltar.at/
------------------------------------------------------------------------------
Stay on top of everything new and different, both inside and
around Java (TM) technology - register by April 22, and save
$200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
300 plus technical and hands-on sessions. Register today.
Use priority code J9JMT32. http://p.sf.net/sfu/p
