Bug#283303: Analysis and possible workaround

[Anton Zinoviev <anton@lml.bas.bg>]

On Wed, Dec 15, 2004 at 10:46:05AM -0500, Jurij Smakov wrote:
> 
> No, this bug does not exists if the swap partition is not the first on the 
> disk. However, as indicated in the bug trail (see my message with subject 
> "[sparc] updated parted_server.c patch"), everything works fine if the 
> changes are committed to disk after filesystem creation using 
> ped_disk_commit_to_dev(). So I believe that the swap partition is 
> positioned correctly (not overwriting the partition table), it is just 
> that changes do not get flushed to disk properly.

If you write the partition table after the creation of the swap then it
depends only on random events whether the partition table will be
overwriten or not.  Especialy if the computer has enough RAM memory then
the partition table will not be overwriten for quiet a long time.

Anton Zinoviev