On Mon, 2002-04-22 at 10:36, Simon Matter wrote:
> You did not tell us which kernel you're on.
Sorry. I think many variables changed at once (hence never reporting
it). Personally, I view it as my own stupid fault.
Factors that changed:
1) went from sane disk layout to dumb disk layout
2) went from RH 7.1 to RH 7.2
3) went from 2.4.3 to 2.4.9-21 (and onto 31) (Both of these 2.4.9's are
your excellent contrib kernels which I also use on my real machines)
> Anyway, if it's the "remount readonly bug" I could explain it like this:
> Before you had /home on a separate partition and this one was unmounted
> on reboot and therefore all data was cleanly written to the disk.
> Now your /home is on the / partition and therefore on reboot your /home
> is remounted readonly and this could leave to data loss if your kernel
> is affected by this bug. Just put a 'sleep 40' in the halt script before
> root gets remounted readonly and try it.
This sounds perfectly feasible. I'll do some tests to see.