double/erraneous fsck bug

sofar sofar at foo-projects.org
Sat Mar 18 02:25:44 UTC 2006


Fixed in moonbase/e2fsprogs. Update your moonbase and relin e2fsprogs to get the new init.d/mount script.

Auke


On Fri, 17 Mar 2006 15:59:12 +0000, sofar <sofar at foo-projects.org> wrote:
> 
> correct. I'll commit a bugfix later today.
> 
> Auke
> 
> 
> On Fri, 17 Mar 2006 14:25:26 +0100, "samuel verstraete"
> <samuel.verstraete at gmail.com> wrote:
>> well i confirm it *doesn't* work on my machine...
>> 
>> i have the latest mount script my fstab entries are as default as they
>> can be and i just rebooted my machine and *all* my disks were checked
>> 3 times...
>> 
>> no fun!
>> 
>> /dev/sda2       /              reiserfs              defaults         0
> 1
>> /dev/sda1       /boot      ext2                     defaults        0  2
>> /dev/sda3       none      swap                  sw                 0  0
>> /dev/sda4       /home    reiserfs              defaults        0 2
>> 
>> 
>> so after a bit of experementing...
>> 
>> 
>> mount script behaves normally (but my /home partition doesn't get
> checked)
>> /dev/sda2       /       reiserfs         defaults        0 1
>> /dev/sda1       /boot   ext2             defaults        0 2
>> /dev/sda3       none    swap             sw              0 0
>> /dev/sda4       /home   reiserfs         defaults        0 0
>> 
>> mount script will check sda1 and sda4 2 times
>> /dev/sda2       /       reiserfs         defaults          0 1
>> /dev/sda1       /boot   ext2             defaults        0 2
>> /dev/sda3       none    swap             sw              0 0
>> /dev/sda4       /home   reiserfs         defaults    0 2
>> 
>> mount script will check sda1,sda2 and sda4 3 times
>> /dev/sda2       /       reiserfs         defaults          0 1
>> /dev/sda1       /boot   ext2             defaults         0 1
>> /dev/sda3       none    swap             sw              0 0
>> /dev/sda4       /home   reiserfs         defaults    0 1
>> 
>> a part from man fstab:
>> The  sixth  field,  (fs_passno),  is  used  by  the fsck(8) program to
>> determine the order in which       filesystem checks are done at
>> reboot  time.   The  root  filesystem  should  be  specified  with  a 
>>      fs_passno  of  1,  and  other filesystems should have a fs_passno
>> of 2.  Filesystems within a drive will be checked sequentially, but
>> filesystems on different drives will be checked at the same  time  to 
>> utilize  parallelism  available  in the hardware.  If the sixth field
>> is not present or zero, a       value of zero is returned and fsck
>> will assume that the filesystem does not need to be checked.
>> 
>> so although 1 1 1 wouldn't be a prefered method to specify the order
>> of checking it should NOT result in checking all filesystem 3 times...
>> 
>> gr,S.
>> 
>> 
>> 
>> 
>> On 3/2/06, sofar <sofar at foo-projects.org> wrote:
>>>
>>> people who have NO systems setup for fsck-ing are hit with this bug.
> The
>>> proper code is a lot simpler and I just committed this to moonbase:
>>>
>>> sed 's/#.*$//g' /etc/fstab | awk '($6>0){print$6}' | sort -n
>>>
>>> in any case, please confirm that the fix works for you by:
>>>
>>> - waiting 10 minutes from now (until a new moonbase is available), and
>> then
>>> - lin moonbase
>>> - lin e2fsprogs
>>> - reboot
>>>
>>> Auke
>>>
>>>
>>> On Thu, 2 Mar 2006 17:43:51 +0000, sofar <sofar at foo-projects.org>
> wrote:
>>> >
>>> >
>>> > On Thu, 2 Mar 2006 10:02:51 -0700, "Alex Hunsaker"
> <badalex at gmail.com>
>>> > wrote:
>>> >> On 3/2/06, Sebastian Kärgel <mailing at wodkahexe.de> wrote:
>>> >>> On Thu, 2 Mar 2006 17:09:34 +0100
>>> >>> samuel verstraete <samuel.verstraete at gmail.com> wrote:
>>> >>>
>>> >>> >
>>> >>> > i might sometimes say very stupid things... and my bash skillz
>> suck
>>> >>> > monkeyballs but that specific thing i checked myself already...
>> and
>>> >>> > they are not listed twice ;)
>>> >>> > gr,S.
>>> >>> >
>>> >>>
>>> >>> Same problem here with /dev/hda1 and /dev/hda2
>>> >>> You're not alone ;)
>>> >>
>>> >> I fixed it with this patch
>>> >
>>> > this patch below makes no sense, unless you have a line with a
>> non-number
>>> > in the fs_passno field.
>>> >
>>> > Can you execute the code snipplet below and paste me the output?
> Also,
>>> > please send me a FULL copy of your fstab
>>> >
>>> > sed 's/#.*$//g' /etc/fstab | awk '{if($6>0){print$1,$6}}' | sort -n
>>> >
>>> >
>>> > Auke
>>> >
>>> >>
>>> >> --- mount.old   2006-03-02 10:01:24.000000000 -0700
>>> >> +++ mount       2006-03-02 10:01:55.000000000 -0700
>>> >> @@ -214,7 +214,7 @@ start()
>>> >>                 fi
>>> >>
>>> >>                 # check filesystems
>>> >> -               FSCKLEVELS=$(sed 's/#.*$//g' /etc/fstab | tr -s
>>> >> '[:blank:]' | awk '{print $6}' | sort -n | grep -v -w 0)
>>> >> +               FSCKLEVELS=$(sed 's/#.*$//g' /etc/fstab | tr -s
>>> >> '[:blank:]' | awk '{print $6}' | sort -n | grep -v -w 0 | egrep
>>> >> '[0-9]+')
>>> >>                 if [ -n "$FSCKLEVELS" ]; then
>>> >>                         echo "Checking file systems:"
>>> >>                         for FSCKLEVEL in $FSCKLEVELS; do
>>> >>
>>> >>> _______________________________________________
>>> >>> Lunar mailing list
>>> >>> Lunar at lunar-linux.org
>>> >>> http://foo-projects.org/mailman/listinfo/lunar
>>> >>>
>>> >> _______________________________________________
>>> >> Lunar mailing list
>>> >> Lunar at lunar-linux.org
>>> >> http://foo-projects.org/mailman/listinfo/lunar
>>> >
>>> > _______________________________________________
>>> > Lunar mailing list
>>> > Lunar at lunar-linux.org
>>> > http://foo-projects.org/mailman/listinfo/lunar
>>> >
>>>
>>> _______________________________________________
>>> Lunar mailing list
>>> Lunar at lunar-linux.org
>>> http://foo-projects.org/mailman/listinfo/lunar
>>>
>> _______________________________________________
>> Lunar mailing list
>> Lunar at lunar-linux.org
>> http://foo-projects.org/mailman/listinfo/lunar
> 
> _______________________________________________
> Lunar mailing list
> Lunar at lunar-linux.org
> http://foo-projects.org/mailman/listinfo/lunar
> 



More information about the Lunar mailing list