Re: System Dorked -- Help!
----- Original Message -----
From: Joel Rees <joel.rees@gmail.com>
Date: Thursday, October 22, 2015 2:34 am
Subject: Re: System Dorked -- Help!
To: d_baron@012.net.il
Cc: debian-user <debian-user@lists.debian.org>
> On Wed, Oct 21, 2015 at 10:32 PM, <d_baron@012.net.il> wrote:
> >
> > ----- Original Message -----
> > From: d_baron@012.net.il
> > Date: Wednesday, October 21, 2015 11:18 am
> > Subject: Re: System Dorked -- Help!
> > To: debian-user@lists.debian.org
> >
> >> ----- Original Message -----
> >> From: d_baron@012.net.il
> >> Date: Wednesday, October 21, 2015 10:24 am
> >> Subject: System Dorked -- Help!
> >> To: debian-user@lists.debian.org
> >
> >> > After being warned of impending failure of my oldie but
> goodie 80gig
> >> > disk -- I had placed my root directory there because the
> installation's>> > partition was too small and did this
> successfully -- I moved the root
> >> > directory to a partition on another disk, edited files, ran
> lilo, seemingly
> >> > successfully, and voile: No boot. Get 99's or nothing at all.
>
> Moved or copied? Did the impending actually fail?
cp -ax oldrootpartition newrootpartition
> When I do something like this, I generally copy and make sure the
> thing works before abandoning the old boot disk.
Yeah, should have had lilo place the mbr to a different disk. Too late now.
> Is LVM involved? or RAID? Do you have a second disk controller
> off the
> motherboard and end up using both controllers so that something thinks
> you were trying to use RAID?
>
No LVM, no RAID
> > Going into a live Debian 7,
>
> I'm guessing that's Kali?
64-bit Debian 7 live.
>
> > I mounted the old and new partitions, copied the
>
> At this point, I'm completely lost about where your boot
> partition and
> other partitions are. Sorry.
>
Partition contents successfully copied before hand. I had neglected to copy the edited fstab and lilo.conf so copied these over in the live session.
> > modified files (which had been done on the old version :-( )
> and chroot and
> > tried to re-run lilo. Segmentation fault. I had been running
> an up-to-date
> > Sid so maybe that is the problem. The lilo is on the
> partition, not on the
> > live distro.
>
> Yeah, but which partition?
>
chroot was to new root partition containing /boot directory and /etc/lilo.conf
The lilo here would not run, segfault.
> >> > Fact is, with certain combination of cabling,
>
> I assume you are using parallel ATA here like my old box?
>
> master/slave and position on those cables is always a bit of a mess.
>
> >> > I had the bootloader
> >> > actually work, load the initrd, and start up, but the new
> root was not
> >> > connected
>
> not connected to what?
I had two of the three disks plugged in. The boot disk was, the root disk was not. Forget about this.
>
> >> > so could not proceed. So what can I do about this?
> >> >
> >
> > More:
> > I installed to the live distro its lilo
>
> I'm having a real problem parsing that. You installed a live distro's
> lilo to the live distro?
>
Yes, I guess from Debian 7, stable?
> Is the live distro on a USB drive or something?
Live CD. It, of course, installs to memory.
>
> > and ran from command line specifying
> > the configuration file and map.
>
> Ran what on what command line?
sudo lilo -C pathtonewextlilo.conf -M pathtonewmapfile
or
sudo lilo -r newrootpartition
>
> > This is what I get:
> > Fatal: Trying to map files from unnamed device 0x0012
> (NFS/RAID mirror down?)
> >
>
> So, I'm not the only one having trouble figuring out what you are
> doing where. ;-)
>
> >> Ok, no segment fault. If I use the -q option, it will display
> my boot
> >> choices.
>
> So is this now in the chroot or booting up in lilo on which device?
>
sudo lilo -r newrootpartition
does the change root and runs the installed lilo there.
> > So, what's next?
> >
> >>
> >
> > And more::
> >
> > Found by Googling that I can bind /proc, /dev to chroot and
> then the
> > installed lilo will run! Same results with or without /proc
> binding. /dev
> > binding got rid of the original error.
>
> In the chroot or from BIOS?
sudo mount --bind newrootpartion/dev /dev
>
> > Boots up with a load of errors but I am now in a machine
> called "Kali Linux"
>
> Is Kali the live distro you are using?
NO, but apparently, somewhere on it, is is.
>
> > I cannot log on to anything. Errors too numerous and fast to
> see. ACPI, does
> > not work on this system anyway. Login service failed to start. Etc.
> >
> > So maybe!! almost there. What else need I do get back to where
> I was?
> > Something else to bind?
> > Copy to restore from the original root directory?
> >
> > (Possibly the "user" login from the live distro might work
> (password?), and
> > I could recreate previous users if need be but that does not
> touch the other
> > errors.)
>
> (Talk about chewing gum and bailing wire.)
>
> First, what was your original partition layout? What was the boot
> disk, where were your other partitions?
Original partitions on one 1 terra disk but their sizes, as per the Debian installation, are ridiculous.
Early on, moved the root partition to the older 80 gigger, went through this process without needing a rescue session and it worked.
The 1 terra SATA remains to boot disk. So get a warning from lilo about differing disks.
I have another 1 terra SATA that has some dorked sectors, why I did that damned installation in the first place. I repartitioned that so that nothing is on the problem area and moved the root partition to a partition on that disk. That is where I stand now. Did not work smoothly this time around, huh?
>
> Can you reconstruct it, even though you may be risking booting
> on a
> failing disk?
I could go back to the original fstab and lilo.conf but would be running lilo off this in the live session. Likely have the same problem again. Disk won't fail sooo fast.
Reply to: