Hi,
I am working on an initramfs-generator (µgRD) focused on stability. In
discussion with the main dev of the project, the question was raised if
our initialization code for LVM would cause issues if we then resume a
prior running system using those same volumes.
Specifically we run
vgchange -ay
vgscan --mknodes
and would then resume, potentially from swap located inside volumes or
potentially elsewhere.
If these two commands could change the state of the disk in any way,
there would be potential that this disparity between disk state and the
state of the then resumed systems cache and kernel could spiral into
serious issues, in the worst case data loss.
Is there any chance of the disk state being changed by those commands?
--
redjard
