Re: 13.3R's installworld killed system--please help!

Ronald Klop <ronald-lists@klop.ws> wrote:

> How big is your pool? Could it be that installworld installed files above > 2TB or some other old BIOS limit so it is unable to read it. 

     I hadn't thought of that, but according to the partition map backups, the
"system" pool is just a hair smaller than 344 GB.  Each of the drives is a bit
smaller than 1 TB.
>
> Sorry if I talk about something that was already discussed. I don?t have the full thread anymore. 
>
     No, it wasn't discussed, so thank you for the input.
     My tower system is still unbootable after all this time.  I'm down to
grasping at straws.  A family member is checking the machines at work to see
whether any still have a usable disk burner in case I need to trash the
contents of one of the boot drives by booting an installer disk.  This would
be a most unpleasant option entailing a huge amount of work and very likely the
purchase of another storage device I can't afford at the moment.
     I'm also considering running the standalone installer from a thumb drive
on the laptop with a boot drive from the tower attached via the USB docking
station.  This would also result in that abominable amount of work.  Neither
this nor the first option is good, but the boot code the standalone installer
installs has worked fine on the laptop's SSD to load 13.3-RELEASE many times
so far, whereas installworld made the system unbootable in a still unknown
manner.
     A third option is to run the installer from a thumb drive on the laptop
in rescue mode to reinstall boot code onto a tower boot drive.  If that fails,
I've considered downloading the last available image of 12.4-RELEASE and
putting that onto a thumb drive, attaching the tower's boot drives to the
laptop, running the 12.4-RELEASE standalone installer from the thumb drive in
rescue mode to install *its* boot code onto the tower's drives, importing the
"system" pool long enough to do a "zfs rollback" of everything to the final
snapshot I made immediately before running installkernel, and putting those
drives back into the tower to try it out.
     Lastly, would be using the 12.4-RELEASE-p2 installer to go through all
the equivalent work and hassle and potential storage device expense by
reinstalling 12.4-RELEASE-p2 from scratch and starting over just to get back to
an approximation of where my system was before I ran installkernel from my
source build of 13.3-RELEASE-p1. :-(  But at least that should give me a
running system again.
     While I've been typing this reply, the aforementioned family member has
reported that her machine at work does have a disk burner.  I doubt it has
been used in a long time, so I don't yet know whether it still works.
     I wonder when the last release was that the BIOS-based boot code was
actually tested on a machine that uses BIOS-based booting.  This leads to
another nasty thought, namely, I don't recall when the last time was that I
installed the boot code onto the boot drives before I got into the current
mess.  It might not have been since I first switched to a ZFS install to begin
using a root-on-ZFS system many years ago, in which case my wondering about
when the developers last tested a BIOS boot becomes somewhat frightening in
that I would have potentially many installer images archived at freebsd.org to
try out.  My pools are all up to date as of 12.4-RELEASE, so the older
releases' boot code versions might not be able to work with the "system" pool
in its current state either.
     Oh, well.  Once again, if anyone seeing this note has any other ideas or
more options to try, please share them.

					Scott Bennett