Reboot loop resolved (I hope)

One thing leads to another…

There are a hand-full of issues – most minor – I have been working my way through. A big one, though, was the reboot loop I mentioned in a couple of posts.

Re-re-re-re-re-re-reboot

For some reason, my system had frequently been getting caught in a re-boot loop, where the boot process would partially complete, the screen would blank, and the (Gigabyte P35-DS4) BIOS screen would appear. This got to the point where I could only boot maybe 1 in 10-12 times. As you might imagine this was frustrating, I posted into the Ubuntu Forums on the matter.

The first thing to do was to try to figure out where in the boot process errors were occurring. This lead me down a few roads:

1. Boot logging. Apparently, on many linux systems /var/log/boot contains the boot log. However, in Ubuntu, /var/log/dmseg is used instead. Admittedly, I do not understand why or if this matters.

In any case, this was of little use, because I couldn’t find the “failed” boots here, only the successful ones.

2. Splash screens. Next step, then was to try to find some way to watch the boot process as it happened. Now, I know that Ubuntu by default shows a “Loading” screen, but it has never worked on my machine – the monitor just blanks during that part. And, I also know that it is possible to watch the whole boot process, because that’s what happens when you choose “recovery mode”.

As it turns out, there are a couple of options here. One is to edit the options at boot time from GRUB, and the other is to edit /boot/grub/menu.lst. By default, the options there are ro quiet splash. Removing quiet splash gets you lots of command line output goodness while booting.

3. Finding the error. As I watched the output for multiple boot (many multiple boots!), I always noticed a long pause and then output like this:

Booting processor 1/4 APIC 0x1...
Not responding.
Inquiring remote APIC #1...
... APIC #1 ID: failed
... APIC #1 VERSION: failed
... APIC #1 SPIV: failed
SMP alternatives: switching to SMP code
Booting processor 2/3 APIC 0x3
Not responding.

This would repeat for each processor, and although the boot process would continue a bit, I was pretty sure this was a problem, especially since a successful boot would look like this:

[ 33.734565] Booting processor 1/4 APIC 0x3
[ 33.744901] Initializing CPU#1
[ 33.822320] Calibrating delay using timer specific routine.. 4800.04 BogoMIPS (lpj=9600089)
[ 33.822326] CPU: L1 I cache: 32K, L1 D cache: 32K
[ 33.822328] CPU: L2 cache: 4096K
[ 33.822330] CPU 1/3 -> Node 0
[ 33.822331] CPU: Physical Processor ID: 0
[ 33.822332] CPU: Processor Core ID: 3
[ 33.822337] CPU1: Thermal monitoring enabled (TM2)
[ 33.822752] Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz stepping 0b
[ 33.822849] checking TSC synchronization [CPU#0 -> CPU#1]: passed.
[ 33.846318] SMP alternatives: switching to SMP code
[ 33.846338] Booting processor 2/4 APIC 0x2

(As an aside, I eventually used my cell phone to take pictures of the screen so I could make sure I was copying things right.)

I searched a lot, and found lots of suggestions along the lines of adding various options to the boot in GRUB. These, and various combinations of these were:

noapic
noacip
acip=off
irqpoll

I also found a couple of people that mentioned disabling USB keyboard and mouse support in the BIOS worked for them.

Now, none of the boot options worked for me, and my BIOS had USB keyboard and USB mouse both disabled by default. I did notice, though, that disabling all USB support in the BIOS seemed to work. Of course, this meant I had to go steal my wife’s PS/2 keyboard, but at least I was making progess!

Finally, I noticed a “Disable Legacy USB Support” in my BIOS. I disabled that, and it seems to have resolved the problem. I’ve rebooted a half-dozen times since then with no problems, so I’m hoping all is well.

Where you at!?

Now I am still unable to get the splash screen working, despite trying the information in this ubunutu forums thread, which apparently works for everyone but me. This is a minor thing, because I really like watching the boot process (I dislike splash screens “hiding” information in general) – but I have to be honest it bothers me because it’s something that isn’t working and it shows how little I know about my own system.