Re: Swap encryption (via LUKS) and Hibernation (disk suspend)
On Saturday 05 May 2007 15:31, bob-at-proulx.com (Bob Proulx) |debian_laptop|
wrote:
> But suspend to ram is not quite so good on my T42. While I can
> demonstrate suspend to ram working I cannot do it reliably 100% of the
> time. This is not related to the encrypted partitions. On the T43p
> (gone to a friend) both types of suspend seem to be quite reliable.
> But unfortunately for me I have not quite converged on a good recipe
> for suspend to ram to be 100% reliable. It sometimes works and
> sometimes fails to resume.
>
> I am still collecting information about it and will eventually be
> posting my own questions to the list about it. But in the meantime I
> have been using suspend to disk with everything encrypted and that has
> not yet failed me.
Sounds great. For me, suspend to ram failure is not very significant. If it
should happen, the only thing that may happen is that any unsaved documents
will be gone and programs will no longer be as they were with suspend to ram.
That is, it would be as if I were to pull the plug from the system.
A suspend to disk failure is much worse because if the swap partition data is
currupted before the laptop resumes, the resuming of the laptop can corrupt
the data on the hard drive. This does not happen to suspend to ram because
either ram resumes, fails or is wiped upon reboot (and therefore fails). This
is why you should never suspend to disk and then try booting from another
kernel (such as booting from a LiveCD) or device. In the LiveCD example, the
LiveCD will boot, mount your swap and use it as its own corrupting the data
that was saved during suspend to disk. When the computer is booted and tries
to resume the suspend to disk from swap, all hell will break loose. This is
my understanding after reading documentation on the different suspend
methods.
Reply to: