Boot works, Reboot doesn't

I have an older Laptop HP ProBook running Linux Mint Debian Edition (LMDE), normally without any issue. The system starts after powering on as expected. But whenever I invoke a restart, e.g. after installing kernel updates, the system hangs with a “Boot device not found” message.
What could be the origin of this issue? How to investigate this?

I’ve to confess that I applied some hardware changes to this system.

  • The original HDD was replaced by a 1TB SSD.
  • The RAM size was increased to 8 GB.

The SSD cannot be the issue because it was the same before with the original HDD in use.

Any thoughts?

3 Likes

Couple of things to start to look at.

Cmos battery is my usual first stop on these issues… boot into the bios (f2, f10, esc ? Could be any of these) check the date and time is correct, usually first point . Depending on the make and model some cmos are easy to change (cr2032) but others are soldered onto the motherboard or connected via a short cable and inside a package.

Next step would be check the startup
When you boot into lmde you get a quick screen display with options to press to arrive at advanced menu, usually 30 seconds, and usually esc key.

This should take you to a short menu where you can run fsck (file system check)
Try running that option
It will ask to confirm or fix errors, select yes to these

There are a couple of other options on this menu to select, sorry not at my desk to explain what each one does, doing this from memory.
You can select any as they all help boot issues.

Once finished, restart your system.

Does that help ?

4 Likes

Thank you, @callpaul.eu. I’ve checked all your suggestions in advance, with no success. The system time is kept, the FS of the boot drive is intact; otherwise the cold boot would fail as well, I assume.

3 Likes

If you try booting from a usb with mint does that cause 5
The same issues ?

When you go into advanced can you see older kernals, if you go back one does that help ?

Is your system up to date with all the updates done ?

2 Likes

Yes, no and yes.

Booted from a life system to check the FS. Fine.
The issue has survived several (>5) kernel updates.
I always keep my systems up to date, at least with the security updates.

EDIT:
It’s not LMDE related. I had Archlinux installed earlier, to try something. Same.

3 Likes

That would suggest more of a hardware related issue.

What are your settings in the bios related to secure boot ?
Something runs in my mind about hp laptops and a similar issue, but not been able to find my notes as to why just remember a bios secure boot being involved

5 Likes

Paul’s on the right track. Secure boot should definitely be disabled, but check also boot devices and boot orders to make sure they’re accurate. I have a Toshiba laptop that required the boot order be edited to put USB first so I could install Devuan Peppermint after I had flushed Windows.

5 Likes

I had to install Debian on a HP micro-machine, whatever it is really called I don’t know.
It had an 8th gen CPU, looked like an intel NUC. It was intended to be a headless server for a friend.
That one had serious problems with reboot, when I initiated reboot from command line, it shut down the system, but never rebooted, it stuck at some point, I saw only blinking cursor on a black screen.
After I updated the BIOS (UEFI), everything worked like a charm.
Maybe a BIOS update will solve your problem too.

7 Likes

I can remember that there was a difference between ‘cold boot’ and ‘soft boot’ on old DOS and CP/M systems. Cold boot does a POST and reinitializes all the hardware, and clears memory. There were situations where firmware loaded into hardware became corrupted while running and required a cold boot to recover.

I dont know if this is relevant to your computer today, but it might be a place to start looking.
“boot device not found” has an obvious meaning… it could not find the disk… so the disk did not respond to a probe … so the disk controller’s firmware may not have been in the correct state.?.. what could corrupt it?.. maybe malware, maybe the wrong driver.

Other thoughts:

  • It would not hurt to run a memcheck.
  • Check your logs after reboot attempt
  • Is there an option in your BIOS to set the disk detection delay? Make it larger… the dusk may not be ready when the BIOS tries to probe it.
5 Likes

No problem so far.

The reboot command seems to disconnect or turn off the drive. In fact, it isn’t longer listed as a selectable boot device. This isn’t a timing issue, the drive remains in its nonexistent state.

1 Like

What reboot should do is sync the disks, then unmount the filesystems… before initating the boot.
The disk isnt seen by the BIOS… hence it is not listed as a boot device.

I wonder if the sync is hanging or the unmount fails ?
Is it any different if you manually sync and unmount all filesystems before doing reboot?.. Hang on that may not be sensible… I dont think you can do umount /

How do you know it is not a timing issue?
The disk has a lot more time to initialize during a cold boot, because the POST takes time. In a reboot you go straight into the BIOS… the disk may not be ready to receive the probe from the BIOS, so the BIOS will not detect it.

Final left field suggsstion:
Try configuring the drive as hot-pluggable in the BIOS
That will change the way the drive is detected.

2 Likes

When you have booted your system and everything is running ok

Please can you look at
Menu, accessoires, disks
This should display all disks attached
Select and screen shot each partition so we can see the results

Or just run
fdisk -l

and paste here the output.

But the “boot device not found” seems to me a BIOS message too, @nevj exactly pointed this out. So I think the problem rises before the system boots (well, for the second time).

Edit:
https://forums.linuxmint.com/viewtopic.php?t=343352

If I get it right, similar symptoms there were healed by BIOS update.

I then updated the BIOS from V05 to the latest V14. Restart and sudo reboot operate as they should.

@abu , can you check wether your MoBo has a BIOS update available?

4 Likes

For the completeness:

Disk /dev/sda: 931.51 GiB, 1000204886016 bytes, 1953525168 sectors
Disk model: Samsung SSD 870 
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x5fcec6ec

Device     Boot      Start        End    Sectors   Size Id Type
/dev/sda1             3906   16800781   16796876     8G 82 Linux swap / Solaris
/dev/sda2  *      16801792   83910655   67108864    32G 83 Linux
/dev/sda3         83910656 1886412799 1802502144 859.5G 83 Linux
/dev/sda4       1886412800 1953521663   67108864    32G 83 Linux

I’m sure, that the disk is fully OK.

It’s more an assumption. If I increase the time the very first BIOS screen (Hit ESC…) is shown, the drive should have enough time to get ready/online, but it doesn’t.

If I reboot into a live system, then the drive will become ready there, even if it wasn’t at the beginning.

Sounds reasonable, but how to achieve this?

I’ll have a look.

1 Like

I really don’t like this! What I have installed:

Mobo: Hewlett-Packard model: 1846 v: KBC Version 58.1D
serial: PCWBF1A2F3W60O BIOS: Hewlett-Packard v: 68IRR Ver. F.32
date: 11/06/2012

The only update available at HP:

date: 07/05/2012 Ver: F.01 Provides the initial BIOS release.

So I’ll keep my fingers off.

1 Like

Wish I had thought of that… but I am a gui man !

Two other thoughts

Linux minu offers a system information option , usually bottom right of screen looks like a clipboard. Can you check that and the info that offers.

I notice that the disk is showing msdos !
Worth formatting to a linux type (ext4) and re installing the system ?

I have had issues with drives in the wrong format showing strange errors similar to this related to speed. In principal linux works with dos drives, but after a certain size it goes strange, experiance says around 1tb, but not logical.

Nope. All my partitions, except swap, are formatted ext4. But the partitioning is MBR, not UEFI.

1 Like

My mistake I read the info you sent and say disklabel type as dos without checking lower down where it clearly shows linux format.

Except for the bios as suggested …

You said you had another linux system on before and it did the same ?
Did you install over that or wipe, format, install ?

There should be a setting in the BIOS.
You need to change something, just to see if we can get some more information.

Reboot is very difficult to debug. You cant use strace, you cant try to replicate it stepwise. All you have is the logs ( look at dmesg) or changing something is the BIOS.

1 Like

Let me summarize the life of this system:

  1. Win7
  2. Win10
  3. LMDE3
  4. LMDE4
  5. LMDE5
  6. LMDE6
  7. Archlinux
  8. LMDE6
  • LMDEs usually changed with dist-upgrade.
  • Hardware upgrade I mentioned above, approx. between 3 and 5.
2 Likes