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

Re: partman problems (sparc and elsewhere)

On Mon, Nov 29, 2004 at 07:22:09PM +0200, Anton Zinoviev wrote:
> On Thu, Nov 25, 2004 at 01:00:51PM -0800, Blars Blarson wrote:
> > 
> > Looking into the partman bugs reports, I think all of these bugs are
> > symptoms of the same partman bug:
> In my opinion most of these bugs are different.
> > My summery is that partman has information in memory on what the
> > desired partitioning is, and information on disk.  It can write
> > information to the disk without updating all of its in memory
> > information to match, and gets itself into an inconsistant state.
> It is possible that libparted does this when it uses sun disk labels
> (but I don't know the internals of libparted).  Partman doesn't have its
> information on the disk.  However there is the following bug in partman:
> sometimes it considers the partition tables in its own data structures
> unchanged and misses to write them to the disk.  The result is that when
> the type of some partition doesn't correspond to the file system in it
> partman may miss to correct the partition type and latter grub fails.

BTW, i also believe that this is a problem on pegasos. Doing a
ped_partition_set_system each time we put a filesystem on a partition may be
the thing to do, do you do this ?

Example, on pegasos, i was going to install an ext3 filesystem in the old swap
partition. The old partition was marked as type swap, and this was not
changed, and thus the ext3 filesystem was not readable from the firmware,
which thought it contained swap, and since there are no driver for this, it

Not sure if it is a general failure to do a ped_partition_set_system, or a
instance of the bug you mention above ? 


Sven Luther

Reply to: