Need help debugging system lockup during boot

I have an HP Envy box core I7 cpu, Nvidia 630 or Intel video,16 GB ram
that freezes during boot of any OS I have.

Running the HP diagnostics at bootup gives a clean bill of health to
memory and all internal hardware as well as extended tests on the disk.
Both video cards work fine with the HP diags on the hard disk.

The freezes are on something down into the boot with no error message. I
have 4 bootable systems on the main drive: Win 10, Tumbleweed, oS 42.3,
and oS 13.2. Each os freezes at the same point into the boot on every
attempt. Booting live versions of Tumbleweed and oS 42.3 results in
freezes at the same message point as the respective systems from the hard
drive so I’m pretty sure that some component is crapping out when
activated but I have no clue what the bootup is trying when the crash
happens.

What I really want to know is: do the oS versions store the boot log to
disk as the boot progresses and if so, how can I access that log? I am
able to pull the drive and access it from another machine and being able
to examine the log of the failed boots would sure help seeing just what
action is crashing the machine. Any info would be greatly appreciated.

That sounds suspiciously like bad ram.

The last time that I had a memory problem, the fix was:

  • pull out a memory module (simm);
  • plug it back in.

The trick is to do that, one at a time, for each simm. It is called “reseating”. And that was all it took.

On Sat, 26 Aug 2017 04:16:02 +0000, nrickert wrote:

> Fraser_Bell;2835626 Wrote:
>> That sounds suspiciously like bad ram.
>
> The last time that I had a memory problem, the fix was:
>
> - pull out a memory module (simm);
> - plug it back in.
>
>
> The trick is to do that, one at a time, for each simm. It is called
> “reseating”. And that was all it took.

Given that 3 different memtest runs had produced no errors overnight, I
was not surprised when re-seating changed nothing. With 2 sticks, I
played musical chairs with them with no change to the boot failure points
as indicated by the progress messages up to failure.

The last thing I tried was to pull all the removable hardware I could one
piece at a time - no change. I had begun to suspect the Nvidia card
after booting a rescuetux cd which found and used it even after disabling
it in the BIOS but that showed no change either.

What I really need is a way to capture the entire bootup message stream
in a way that permits examination of the process. I wish I had access to
the old-school debug trick of re-directing the boot stream to an external
terminal but that requires having the debug kernel installed. I’m
leaning toward a hard error in a device (CPU, memory manager, etc) but
without the boot message stream that leaves me with no real option but
part swapping. Not an attractive proposition given the cost of that I7
processor…

One possibility: anybody know of a single-processor (non-SMB) distro I
could use to play? Rescuetux lacks a lot of programs need for debugging
this but it does boot and run.

During boot,
You can typically hit ESC on the keyboard to view the bootlog messages on screen as the boot progresses.
Of course, if your system freezes, then the last boot entries should be frozen on the screen.

If you can boot into emergency mode at the GRUB menu,
From the running system you can view any prior boot log with the following command

journalctl -b -*number* 

So, for instance you can read the prior bootlog with the following

journalctl -b -1

You can also further filter by error levels, so for instance the following will return all critical level errors from the last boot

journalctl -b -1 -p "crit"

Also,
Whenever I’ve re-seated electrical connections, I’ve also cleaned them with a pencil eraser.

Also, consider this…
Laptops are not that easy to pull apart, and components are generally soldered to the systemboard and not removable.
Unless you’re considering using a soldering iron to remove components, I’d recommend instead that you see what a replacement laptop sells for (If your system is more than a couple years old, a used machine will often cost less than $100).
Then, all you’d need to do is simply move your hard drives to the replacement machine.

TSU

On Mon, 28 Aug 2017 15:06:01 +0000, tsu2 wrote:

> From the running system you can view any prior boot log with the
> following command
>
> Code:
> --------------------
> journalctl -b --number-
> --------------------
>
> So, for instance you can read the prior bootlog with the following
>
> Code:
> --------------------
> journalctl -b -1
> --------------------
>
>
> You can also further filter by error levels, so for instance the
> following will return all critical level errors from the last boot
>
> Code:
> --------------------
> journalctl -b -1 -p “crit”
> --------------------
>

That’s exactly what I was looking for. Thank you - although it’s obvious
since you point it out… RTFM time!

>
> Also,
> Whenever I’ve re-seated electrical connections, I’ve also cleaned them
> with a pencil eraser.
>

Goes without saying - but I have found that pencil erasers can be more of
a contamination than a help unless you know for sure it’s clean. I keep
a drafting eraser that I dedicated to this purpose years ago.

> Also, consider this…
> Laptops are not that easy to pull apart, and components are generally
> soldered to the systemboard and not removable.
> Unless you’re considering using a soldering iron to remove components,
> I’d recommend instead that you see what a replacement laptop sells for
> (If your system is more than a couple years old, a used machine will
> often cost less than $100).
> Then, all you’d need to do is simply move your hard drives to the
> replacement machine.
>

Fortunately, this is a relatively new tower with lots of room to work and
big access panels. It’s also a challenge and retirees like myself have a
lot more time than money for what is now a hobby.