[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#637784: installation-reports: Partition disks error: No root file system is defined.



Le Mon, Feb 13, 2012 at 10:06:06AM +0900, Charles Plessy a écrit :
> Le Sun, Feb 12, 2012 at 04:40:38PM +0900, Charles Plessy a écrit :
> > Le Sat, Aug 20, 2011 at 06:26:45PM +0900, Charles Plessy a écrit :
> > > 
> > > Using the shell console, I found that the partition table was the following:
> > > 
> > >   Disk /dev/xvda1: 1073 MB, 1073741824 bytes
> > >   255 heads, 63 sectors/track, 130 cylinders
> > >   Units = cylinders of 16065 * 512 = 8225280 bytes
> > >   Sector size (logical/physical): 512 bytes / 512 bytes
> > >   I/O size (minimum/optimal): 512 bytes / 512 bytes
> > >   Disk identifier: 0x0002dcf2
> > >   
> > >       Device Boot      Start         End      Blocks   Id  System
> > >   /dev/xvda1p1   *           1         118      947803+  83  Linux
> > >   /dev/xvda1p2             119         130       96390    5  Extended
> > >   /dev/xvda1p5             119         130       96358+  82  Linux swap / Solaris
> > > 
> > > However, the device files /dev/xvda1p1, p2 and p5 are not available.  Do you have any idea ?
> 
> I tried to modify the partition table by hand using fdisk, and after writing it
> I see the following error message.
> 
>   WARNING: Re-reading the partition table failed with error 22: Invalid argument.
>   The kernel still uses the old table. The new table will be used at
>   the next reboot or after you run partprobe(8) or kpartx(8)
>   Syncing disks.
> 
> So the scenario is:
> 
>   - Partman partitions the disk,
>   - the kernel is not informed of the changes,
>   - udev does not create new links in /dev,
>   - formatting and installation are impossible.

Hello again,

it looks like /dev/xvda1, where the Xen system is booted, is a special case and
that in contrary to the other devices, like /dev/xvdb, etc., it can not be
modified.

The problem is therefore that part of the toolchain does not recognise in
advance that it is impossible.

Do you think that partman can or should obtain this information by itself, or
that this bug should be reassigned to another package, which would be in charge
of issuing an error message that partman can understand, to make sure that
debian-installer does not assume that the disk has been partitionned
successfully ? 

-- 
Charles Plessy
Tsurumi, Kanagawa, Japan



Reply to: