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

Re: partman-auto limitations on sparc

On Thu, Nov 11, 2004 at 01:16:29AM -0500, Joey Hess wrote:
> On sparc, partman-auto can only partition existing free space, it's not
> capable of repartitioning a whole disk. Indeed, it does not even present
> that as an option, and if it's preseeded to partition a whole disk
> anyway, we see why:
> /bin/perform_recipe: IN: NEW_PARTITION =dev=ide=host0=bus0=target1=lun0=disc primary ext3 0-4303272959 beginning 100000001
> parted_server: OUT: Can't have the end before the start!
> So, what's the story behind this, and can it be fixed? It makes it hard
> to do useful preseeded automatic installs, and so it's blocking me from
> automated sparc install testing. I can work around that by writing an
> expect script to drive partman (creating a new partition table on the
> device and then partitioing its newly free space seems to work), or I can
> try to fix partman, whichever is easier. :-)

As said on irc, i believe that the problem is twofold here, that is there are
two parted bugs/features that influence this : 

  o parted has a bug showing large partitions, du to it using an int and not a
  larger kind of value. I have a patch for this, and it will go into -9.

  o parted doesn't seem to cope well with partitions showing the whole disk,
  since it conflicts with the logic that tests that a given sector can only
  be part of a single partition. Not sure if this is fixed or not though.

Anton, do you have any insight on this ? It may be due to partman's shell
calculations also.


Sven Luther

Reply to: