Call blockdev --rereadpt after sfdisk commands.
authorRichard W.M. Jones <rjones@redhat.com>
Wed, 22 Sep 2010 15:35:52 +0000 (16:35 +0100)
committerRichard W.M. Jones <rjones@redhat.com>
Wed, 22 Sep 2010 15:35:52 +0000 (16:35 +0100)
commit956fc5a3feacc970ea763697bf28fb686c875408
tree678041965c82e898d8852b45ebd520db85f61962
parentf6fda69953ca110104481f1263ec9a96e74dbb88
Call blockdev --rereadpt after sfdisk commands.

On fast machines sfdisk has some sort of race where it
fails to re-read the partition table it has just created
(it's not clear if this is a race in sfdisk, the kernel or
some other component).

This commit works around the problem by calling
blockdev --rereadpt after sfdisk operations, which
experience shows is enough to stop the problem from
happening.
daemon/sfdisk.c