FreeBSD - Automatic reboots on boot attempts - init / getty failure?

Question

$ uname -a
 FreeBSD MACH1 10.2-RELEASE FreeBSD 10.2-RELEASE #0 r286666: Wed Aug 12 19:31:38 UTC 2015 [email protected]:/usr/obj/usr/src/sys/GENERIC i386

On booting, everything seems to run OK up to a point. POST runs as expected; hardware/device probe runs with bright white foreground text fonts to stdout. I can't tell if the probe is completed because of its rapidity, but apparently stalls afterwards as there is no further output, in fact the screen display disappears.

The system at this point or after some irregular delay period will do an automatic reboot, [BIOS being set for auto reboot on 'Power OFF'].

There is no indication of entering the init or getty stage of the bootstrapping process? Perhaps an abort. i.e. no grey text displayed! There is no evidence of a 'PANIC' either.

My mobo caps seems fine, no bulges or leaks! Of course I can't see those in the PSU, although I swapped another in. There is no dmesg to retrieve.

Now, if I boot with the livefs {FreeBSD 10.2 Release Disc} - system boots fine.

mount -t ufs /dev/gpt/gptrootfs /mnt
mount -t devfs /dev /mnt/dev
chroot /mnt
mount -a

Apart from a couple of fscks_ufs everything seems fine and stable.

Shutdown, followed by a power-up and the problem re-asserts itself. My /boot/loader.conf is currently empty no customization & everything should be 'at default", yet - continual reboots.

Returning to the livefs, I have to do a fsck_ufs on all the disks, system and non-system disks:

 The universal precipitating message is

 "mount will invalidate journal contents"

 Then the traditional fsck stuff with a
 ...
 RECOVER ? yes
 **building recover table**
 **Resolving unreferenced inode list**
 **Processing journal entries**
 WRITE CHANGES? yes
 *****File System Marked Clean*****

If I do a normal boot again, then the same auto reboots occur, necessitating an unceremonious shutdown and perhaps dismount; anyway apparently the system becomes dirty. Given that the system did not mount I could have expected this of the system disk but not all disks!

Isn't it safe to assume that since the system is stable with the livefs and all the disks mounted that I can eliminate any consideration of a hardware problem?

Is there some way of figuring out what is really going on here; and what are reasonable possible solutions? Hopefully not a re-installation.

Answer 1

You might try to boot the machine with an argument like "init=/bin/sh" (sorry, I'm not familiar with BSDs (anymore), so you have to find out how to run a shell as init(8)). With extreme care (you are running with privileges that are even higher than root, with less of a safety net than in the regular root account, and in a possibly severely crippled environment due to sundry services not having been initialized) you can rummage around and try to fix the problem. Read up on how to start required services too.

If available, probably an installation CD/DVD/pendrive will offer a "maintenance" option, or boot that one far enough to be able to mount your system for diagnosis.

FreeBSD might even offer "maintenance disks", or perhaps there are third party ones. If available, this would be your best bet.