On Sun, 5 Sep 1999, Luqman Hakim wrote:
> > Luqman Hakim wrote:
> > >
> > > I've reinstal Mandrake 6.0
> > > upgrade the kernel, and initscript
> > > modifying lilo.conf
> > > then run /sbin/lilo
> > > then I test the power failure by pushing the power
> > > button in X (I recently work in X) without any application
> > > running.
> >
> > Pardon? Sorry, but I just can't believe this. Have you any idea of how
> > many files are open and processes are running on your box (run 'top' to
> > find out)? DON'T DO THIS! There is no such 'feature' that prevents damage
> > if you reboot your box right from the middle of operating.
> >
> I think it still safe if my harddisk is not currently read/writing.
> Like other OS like windows will automatically running scandisk or NDD
> Linux automatically run the e2fsck.
Negative, unlike windows on a linux system there is alot more cacheing
involed, type 'free' and check it out, i've never seen less than 18Megs
precached on this machine.
> >
> > > Then I think the problem fixed in upgraded kernel
> > > and initscripts is that my RPM database is still exist.
> > > I can also login at the prompt.
> > > The "mount/unmount [failed]" problem is always
> > > exist.
> >
> >
> > What did you expect? You didn't unmount them! Do you expect Linux to spare
> > some extra electricity cycles for such desasters?
>
> I do unmount them!. it's included in level 6
> the 'shutdown -h now' command should unmount it.
> this is the message:
> Shutting down NFS mountd: [FAILED]
> Shutting down NFS daemon: nfsd: terminating on signal 9
> nfsd: last server exiting [FAILED]
Try and find out why your nfs daemons do not stop, this is the reason your
filesystems are always busy and can not be unmounted..
> >
> > > (Below i include my previous problem
> > > before upgrading the kernel.)
> > >
> > > My friend told me that his SuSE has no problem
> > > with power failure. He test it by pushing the power button
> > > as i do, even many time.
> >
> > He is a *very* lucky guy, then. And the Linux fs can take a blow.
> > Nevertheless it's utterly loony to do this on purpose (sorry). Power
> > failures are worst cases accidents and only UPS can prevent you from
> > damage. Nothing else. (Well, apart from the journaling fs, but this will
> > take a while till it gets implemented in ext2fs)
> >
> In other Linux mailing list, someone just told me that he have Mandrake
> 6.0 on his computer. When he work on RPM (installing files that quite large)
> then eletrical suddenly cut-down. but Mandrake is still okay.
> I have really no idea about this.
>
> Maybe i'm the only one who got this problem.
> oh no, in other mailing list (in my province)
> someone has similiar problem with me
> he done it on his laptop. He told me that maybe the
> /usr/share/config/kdmrc is the primary problem
> it has 0 byte when the power failure happens.
>
> or
> Could it be the partition ?
> I've only set it to 3 part (root, swap, home)
Double check /etc/inittab, for a line that reads
ud::once:/sbin/update
> Now, i've reinstall Mandrake, and set it to 5 partition
> (root, boot, swap, home, usr), it's running well.
> I don't wanna do the power test anymore, i'm tired!
>
>
> Luqman
>
>
--
MandrakeSoft http://www.mandrakesoft.com/
--Axalon
