[mythtv-users] Problem with boot time system lockup on new install of Mythbuntu 12.04

[Craig Huff]

I have a problem that I can't seem to pin down with booting on a newly
installed Mythbuntu 12.04 (MythTV 0.25) system.  It randomly freezes
during boot after starting to report starting up several daemons from
/etc/rc2.d/S*, but before switching to logging in /var/log/*, which is
making it _very_ challenging for me to figure out what the problem is.
 Based on the fact that it it happens at random points in the boot
process (varying numbers of daemon starts reported) and randomly
occurs (often, the system boots successfully for several tries before
one or more failures), it appears to be some kind of race condition,
but I haven't been able to pin it down.

I am alternating between the V0.23 and V0.25 systems by booting from
separate installs on separate disks, so the hardware is the same in
all respects other than the disk from which it is booted.  All the
recordings (*.mpg, etc.) are on other disk drives, leaving only the
O/S, programs, logs, swap partition, and mysql data on the boot drive.
 The V0.23 system is as reliable as one could ask for, hence the long
delay in upgrading to a newer version of Mythbuntu/MythTV.

I thought it might have something to do with my IguanaWorks USB IR
tranceiver daemon, but I completely removed it from the boot process
by disabling the /etc/init.d script and moving the udev rules for it
out of both /etc/udev/rules.d and /lib/udev/rules.d with no change, so
I'm putting them back in at the moment.

Just to cover another possiblity, be aware that the tuners are DTAs
attached to the four inputs of a pair of Hauppauge PVR-500 PCI cards,
and I haven't had any problems with them under V0.23, so I doubt there
is any connection to this problem from that direction.

I tried re-activating an /etc/init/research.conf rule I created a long
time ago that basically started on _any_ *-device-added event.  It
posts the date/time and all environment variables to
/dev/.initramfs/research.log.  At the time I created it, I was
tracking down an issue with detecting devices and when the detections
occurred.  It worked then, but when I tried it on a failing boot
attempt, there was nothing in /etc/.initramfs, not even the
subdirectory .initramfs.  (I checked after a boot crash by rebooting
into the UBCD PartedMagic system and looking at the target boot disk
contents.  While in a terminal window on PartedMagic with the boot
disk mounted, the command:
    find /media/sda1 -mtime -1 -ls | grep "Apr 15"
returned no files found, so the lockup is happening before anything
gets written to disk.  Note that all of the disks had been reported
mounted and fsck'd during the boot process before the lockup, so they
were available before the crash occurred.

I'm running out of ideas on what to try next.  Any suggestions welcomed.

--
Craig.