Solved Boot loop after upgrade 11.1 -> 11.2

Whilst upgrading from 11.1 to 11.2 I entered the last phase -following the instructions here and rebooted with the aim of running /usr/sbin/freebsd-update install and my system goes into a boot loop.

What do I need to do to identify what is happening?
 
The system is crashing on booting, even in single user mode. I tried escape to the loader prompt but had no idea as to what to do.

Eventually I selected Option 5 from the boot menu and was presented with an option to boot using kernel.old which did boot up OK but now I'm back at 11.1.

How do I figure out what is wrong?
 
After discovering the Configure Boot Options, I selected Safe Mode and Verbose, and discovered:-
Fatal trap 12: page fault while in kernel mode

I tried the same disk in various laptops but got the same response each time.
 
Whilst upgrading from 11.1 to 11.2 I entered the last phase -following the instructions here and rebooted with the aim of running /usr/sbin/freebsd-update install and my system goes into a boot loop.
Sorry but I don't understand what this means. What exactly did you do before rebooting?

I get the impression that you ran upgrade, and instead of running install right after it (as instructed by the manual) you decided to reboot and ended up with this mess. If so then it should be obvious what happened here and how to fix this.

But what exactly did you do? You don't give any useful information here.
 
Sorry but I don't understand what this means. What exactly did you do before rebooting?

I get the impression that you ran upgrade, and instead of running install right after it (as instructed by the manual) you decided to reboot and ended up with this mess. If so then it should be obvious what happened here and how to fix this.

But what exactly did you do? You don't give any useful information here.

I followed the instructions here.

Code:
# freebsd-update fetch
# freebsd-update install

# freebsd-update -r 11.2-RELEASE upgrade

# freebsd-update install

# shutdown -r now

Intending to run this on reboot but boot loop prevented it being run:-

# freebsd-update install
 
Sure, but as you can see the install command is mentioned three times and all you told us is that you rebooted in order to run install after that. So how are we supposed to know at which point that was?


I thought I explained exactly what I diid....


The machine should now be restarted with the updated kernel:

# shutdown -r now
Once the system has come back online, restart freebsd-update using the following command. Since the state of the process has been saved, freebsd-updatewill not start from the beginning, but will instead move on to the next phase and remove all old shared libraries and object files.

# freebsd-update install

I was unable to run the last bit because the system did not come up cleanly and gets stuck in a boot loop.

I thought I had explained it clearly previously.
 
I thought I explained exactly what I diid....
Thing is: I do not go on assumptions and "last phase" didn't tell me anything. I'm not trying to be difficult but I've simply experienced way too many issues where "last" meant something completely different to someone than that which others would expect.

Water under the bridge, now that we got this cleared up let's get to business.

First of all; side comment: I don't understand why so many newbies are suddenly commenting on using the vc terminal. It makes no sense because vt has become the default long before 11.2, so it seems a bit weird to suddenly assume that this could be causing issues.

Let's try not to derail this thread and stick with the facts that actually apply to balanga? So far he never shared anything which hints in this direction I might add. Doesn't imply that it cannot be related, but it seems pretty plausible at this time (time of writing).

So it seems that your problem is kernel related, also concluding from the fact that your old kernel works normally.

Can you share what's in your /boot/loader.conf?

Also: if you have the option then I'd definitely encourage you to try and boot the system with a 11.2 rescue setup. Even installation media would be good. This can rule out any issues with the GENERIC kernel vs. your system setup.
 
Can you share what's in your /boot/loader.conf?
vboxdrv_load="YES"
if_urndis_load="YES"
cuse_load="YES"
Actually I just renamed /boot/loader.conf and it boots normally so one or more of the above is causing the problem...

Now.... to remember why they were included...
Also: if you have the option then I'd definitely encourage you to try and boot the system with a 11.2 rescue setup. Even installation media would be good. This can rule out any issues with the GENERIC kernel vs. your system setup.

I don't know anything about a rescue setup but would be interested in learning about it.

Many thanks for the help.
 
My guess is that the vboxdrv driver is the culprit here, there was a change to the kernel API/ABI in 11.2 that is incompatible with some kernel modules built on 11.1.

Boot to single user mode and comment out the vboxdrv_load="YES" line, reboot. Build a new kernel module for the virtualbox guest drivers using the emulators/virtualbox-ose-additions port and re-enable the driver in loader.conf(5).

And for people suggesting random junk like the kern.vty trick, please stop jumping to conclusions before there's enough information of what the actual problem might ,be.
 
I commented out
vboxdrv_load="YES"

and everything is OK, but your advice about booting to single user mode did not work as the system went into its boot loop as usual. What you need to do is select option 5 at the boot menu and boot with the old kernel.
 
Yes, I mixed up single user mode with breaking into the loader prompt at the boot menu. You could have booted the new kernel as well by first unloading everything and then loading the new kernel followed by boot. If you load a kernel at the loader prompt none of the kernel modules are loaded unless you explicitly load them.
 
I updated my system earlier and foolishly forgot to comment out vboxdrv_load="YES". I am now in the same boot loop as described here. Single user mode doesn't help and nor does using the old kernel. Going into the boot loader prompt and doing "unload" doesn't help, neither does "disable-module vboxdrv".

Anybody have any ideas how I can get back in?

Also, does anybody know why this is still an issue so long after 11.2 came out?

EDIT: Fixed it by booting from USB, mounting the filesystem, commenting out the relevant VirtualBox and nvidia lines from /etc/rc.conf and /boot/loader.conf, rebooting and building the relevant ports. I'd still like to know why I am having to do this after all this time?
 
For your information, I just updated to freebsd 11.2 and had the same problem.

I did the following to resolve it:
1. in the boot loader, choose option 3 to get a prompt
2. type disable-module vboxdrv + press enter
3. type boot + press enter

Now the system boots, not loading vboxdrv.

But it still restarted, because I had something of virtualbox in my rc.conf.
Luckily, due to a wrong configuration with encryption, the boot process drops me to a prompt half-way the booting, where I have to enter
zpool import -a

There, I could uncomment the vboxdrv lines in my loader.conf and in rc.conf, using ed.

I must remember to comment the vboxdrv stuff first, when upgrading to 12.0. :)

The solution remains the same: try to edit the loader.conf and uncomment the lines from virtualbox.
 
Update your packages, this issue should be resolved now as 11.1 is EoL and packages are built for 11.2.
 
Back
Top