Bug#400024: mondo 2.20: wrong mountlist

2006-11-25 Thread hugo vanwoerkom

On 11/25/06, Andree Leidenfrost <[EMAIL PROTECTED]> wrote:

[Bruno: If you have anything to add or correct, please let us know.]

Hi Hugo,

Would be great if you could cc the bug report. Other than that, thanks a
lot for sending the additional information through! I think I know what
is going on now: The key point is that all filesystems that are not part
of the mountlist have the 'noauto' flag set. I assume that they are not
mounted when mondoarchive/mindi runs.

mindi's behaviour is actually as designed: Filesystems that have the
'noauto' option set in /etc/fstab and that are NOT mounted will not be
included in the mountlist. So, the mountlist you are getting is actually
correct in that it only contains the filesystems that don't have the
'noauto' option set and assuming that none of the filesystems that have
the 'noauto' option set are mounted. (I have tested this on my end and
it works as expected. It would be great if you could confirm that you
have none of the 'noauto' filesystems mounted when mindi runs, though.)

For the record, the code to explicitly check the above condition went
into mindi in SVN revision 717:
http://trac.mondorescue.org/changeset/717

I don't think that mindi's behaviour is a bug or could be improved
really: mondoarchive (and with it mindi) is not mounting or unmounting
any filesystems of the system it runs on. This is left to the
administrator and, as I believe rightfully so - bad things can happen if
filesystems are willy-nilly mounted and unmounted. So, in your
situation, in order to get everything archived you want to, my
recommendation would be to manually (or using a script) mount the
filesystems in question prior to running mondoarchive. (Alternatively,
in case you don't want to mount them, you might also be able to use '-x'
but that would create rather inflexible dd images of those partitions.)

Also, please note that the mountlist generated is only used for the
restore process. /etc/fstab in a restored system will look exactly the
same as before.

I may have overlooked something, and if so would be keen to find out.
Either way I'd much appreciate your feedback on the above.

Best regards & thanks a lot,
Andree



Thanks a lot Andree. That explains it. The partitions in fstab are
indeed not mounted because if they were mondo would try to back them
up, unless I exclude them with the -E option, if I am not mistaken.
Great job!

Hugo




On Fri, 2006-11-24 at 07:50 -0600, hugo vanwoerkom wrote:
> On 11/23/06, Andree Leidenfrost <[EMAIL PROTECTED]> wrote:
> > reassign 400024 mindi
> >
> > thanks
> >
> >
> > Hello Hugo,
> >
> > Thank you for reporting this problem.
> >
> > It would be great if you could run the following command (as root):
> >
> > mindi --makemountlist /tmp/mount.lst
> >
> > and send me:
> > - the screen output
>
> Script started on Fri 24 Nov 2006 07:27:15 AM CST
> executing /root/.bashrc
> /Fri Nov 24-07:27:15HDC3# mindi --makemountlist /tmp/mount.lst
> lilo.real found; will be used instead of lilo (*grumble* *mutter*)
>
> Your mountlist will look like this:-
>
> Hang on...
> DEVICE  MOUNTPOINT  FORMAT  SIZE (MB)   LABEL
>
> BLKGETSIZE ioctl failed on proc
> /dev/hdc3   /   ext2   7632
> /dev/hda9   swapswap953
> /dev/sda7   /sda7   ext2  38154
> /Fri Nov 24-07:27:21HDC3# exit
> exit
>
> Script done on Fri 24 Nov 2006 07:27:27 AM CST
>
> > - /tmp/mount.lst
>
> /dev/hdc3 / ext2 7815622
> /dev/hda9 swap swap 976712
> /dev/sda7 /sda7 ext2 39070048
>
>
> > - /var/log/mindi
>
> mindi v2.2.0-r881
> i686 architecture detected
> mindi called with the following arguments:
> --makemountlist /tmp/mount.lst
> Start date : Fri Nov 24 07:27:19 CST 2006
> MINDI_LIB = /usr/lib/mindi
> MINDI_SBIN = /usr/sbin
> MINDI_CONF = /etc/mindi
> MONDO_SHARE =
> Found isolinux.bin at /usr/lib/syslinux/isolinux.bin
> lilo.real found; will be used instead of lilo (*grumble* *mutter*)
> Your raw fstab file looks like this:-
> # /etc/fstab: static file system information.
> #
> #
  
> proc/proc   procdefaults  
0   0
> /dev/hdc3   /   ext2defaults,errors=remount-ro
0   1
> /dev/hda9   noneswapsw
0   0
> /dev/hdd  /cdrom  iso9660 ro,user,noauto  
0   0
> /dev/fd0/media/floppy0  autorw,user,noauto
0   0
>
> # 80GB PATA Maxtor disk
>
> /dev/hda1   /hda1   ext2
> defaults,user,noauto,errors=remount-ro0   1
> /dev/hda5   /hda5   ext2
> defaults,user,noauto,errors=remount-ro0   1
> /dev/hda6   /hda6   ext2
> defaults,user,noauto,errors=remount-ro0   1
> /dev/hda7   /hda7   ext2
> defaults,user,noauto,errors=re

Bug#400024: mondo 2.20: wrong mountlist

2006-11-25 Thread Andree Leidenfrost
[Bruno: If you have anything to add or correct, please let us know.]

Hi Hugo,

Would be great if you could cc the bug report. Other than that, thanks a
lot for sending the additional information through! I think I know what
is going on now: The key point is that all filesystems that are not part
of the mountlist have the 'noauto' flag set. I assume that they are not
mounted when mondoarchive/mindi runs.

mindi's behaviour is actually as designed: Filesystems that have the
'noauto' option set in /etc/fstab and that are NOT mounted will not be
included in the mountlist. So, the mountlist you are getting is actually
correct in that it only contains the filesystems that don't have the
'noauto' option set and assuming that none of the filesystems that have
the 'noauto' option set are mounted. (I have tested this on my end and
it works as expected. It would be great if you could confirm that you
have none of the 'noauto' filesystems mounted when mindi runs, though.)

For the record, the code to explicitly check the above condition went
into mindi in SVN revision 717:
http://trac.mondorescue.org/changeset/717

I don't think that mindi's behaviour is a bug or could be improved
really: mondoarchive (and with it mindi) is not mounting or unmounting
any filesystems of the system it runs on. This is left to the
administrator and, as I believe rightfully so - bad things can happen if
filesystems are willy-nilly mounted and unmounted. So, in your
situation, in order to get everything archived you want to, my
recommendation would be to manually (or using a script) mount the
filesystems in question prior to running mondoarchive. (Alternatively,
in case you don't want to mount them, you might also be able to use '-x'
but that would create rather inflexible dd images of those partitions.)

Also, please note that the mountlist generated is only used for the
restore process. /etc/fstab in a restored system will look exactly the
same as before.

I may have overlooked something, and if so would be keen to find out.
Either way I'd much appreciate your feedback on the above.

Best regards & thanks a lot,
Andree


On Fri, 2006-11-24 at 07:50 -0600, hugo vanwoerkom wrote: 
> On 11/23/06, Andree Leidenfrost <[EMAIL PROTECTED]> wrote:
> > reassign 400024 mindi
> >
> > thanks
> >
> >
> > Hello Hugo,
> >
> > Thank you for reporting this problem.
> >
> > It would be great if you could run the following command (as root):
> >
> > mindi --makemountlist /tmp/mount.lst
> >
> > and send me:
> > - the screen output
> 
> Script started on Fri 24 Nov 2006 07:27:15 AM CST
> executing /root/.bashrc
> /Fri Nov 24-07:27:15HDC3# mindi --makemountlist /tmp/mount.lst
> lilo.real found; will be used instead of lilo (*grumble* *mutter*)
> 
> Your mountlist will look like this:-
> 
> Hang on...
> DEVICE  MOUNTPOINT  FORMAT  SIZE (MB)   LABEL
> 
> BLKGETSIZE ioctl failed on proc
> /dev/hdc3   /   ext2   7632
> /dev/hda9   swapswap953
> /dev/sda7   /sda7   ext2  38154
> /Fri Nov 24-07:27:21HDC3# exit
> exit
> 
> Script done on Fri 24 Nov 2006 07:27:27 AM CST
> 
> > - /tmp/mount.lst
> 
> /dev/hdc3 / ext2 7815622
> /dev/hda9 swap swap 976712
> /dev/sda7 /sda7 ext2 39070048
> 
> 
> > - /var/log/mindi
> 
> mindi v2.2.0-r881
> i686 architecture detected
> mindi called with the following arguments:
> --makemountlist /tmp/mount.lst
> Start date : Fri Nov 24 07:27:19 CST 2006
> MINDI_LIB = /usr/lib/mindi
> MINDI_SBIN = /usr/sbin
> MINDI_CONF = /etc/mindi
> MONDO_SHARE =
> Found isolinux.bin at /usr/lib/syslinux/isolinux.bin
> lilo.real found; will be used instead of lilo (*grumble* *mutter*)
> Your raw fstab file looks like this:-
> # /etc/fstab: static file system information.
> #
> #
>   
> proc/proc   procdefaults  
> 0   0
> /dev/hdc3   /   ext2defaults,errors=remount-ro
> 0   1
> /dev/hda9   noneswapsw
> 0   0
> /dev/hdd  /cdrom  iso9660 ro,user,noauto  
> 0   0
> /dev/fd0/media/floppy0  autorw,user,noauto
> 0   0
> 
> # 80GB PATA Maxtor disk
> 
> /dev/hda1   /hda1   ext2
> defaults,user,noauto,errors=remount-ro0   1
> /dev/hda5   /hda5   ext2
> defaults,user,noauto,errors=remount-ro0   1
> /dev/hda6   /hda6   ext2
> defaults,user,noauto,errors=remount-ro0   1
> /dev/hda7   /hda7   ext2
> defaults,user,noauto,errors=remount-ro0   1
> /dev/hda8 /mnt/win_F  vfatuser,noauto,exec,umask=0
> 0   0
> #/dev/hda10   /hda10  ext2auto,user,exec  
> 0   0
> /dev/hda10  /hda10  ext2
> defaults,

Bug#400024: mondo 2.20: wrong mountlist

2006-11-23 Thread Andree Leidenfrost
reassign 400024 mindi

thanks


Hello Hugo,

Thank you for reporting this problem.

It would be great if you could run the following command (as root):

mindi --makemountlist /tmp/mount.lst

and send me:
- the screen output
- /tmp/mount.lst
- /var/log/mindi

Also, unless I am mistaken you are not running a stock Debian kernel but
a custom one. Is there any particular reason for this? Could you retest
with a stock Debian?

Finally, I'd just like to confirm that what you are saying is that if
you downgrade to 2.0.8 and run it on your *current* system, it works ok
but 2.0.9 and 2.2.0 don't.

Best regards,
Andree
-- 
Andree Leidenfrost
@ Debian Developer
Sydney - Australia



signature.asc
Description: This is a digitally signed message part