Solved Lost GUI and SSH on upgrading to 13.2 - advice on how you would proceed

D

dbtrade

TL:dR;
After upgrading to 13.2 STABLE from 12.4, everything was fine for a day.
But now I have lost the GUI (was XFCE4) and cannot SSH in.

Where would you start troubleshooting this? I would be truly grateful for any advice you may have.
For example, try and get SSH working, the desktop working? Try and figure out why these are not working to begin with? .shrc problems?

I love the box, but I am not under the hood with it much. I have a backup of the original os disk prior to the upgrade, but not sure all of that would be the way to go.
What would you do?

I still see my shared storage pools.

Details:
Everything seemed fine after the upgrade, although my bhyve VM which was running an instance of HomeAssistant would no longer boot.

After researching, I tried downgrading a package that was causing others issues. I ran package checks all was good. It didn't fix the issue so I reverted until I could figure something else out. I was restarting between these actions and all seemed fine.

What else was going on, I have always used su for root access, I started to use sudo -i after reading about it in a thread.
At any rate much later in the day, I put a Raspberry Pi on the network to temporarily run Home assistant. I had to update it using a backup from when the instance was running in the VM. There was no network clash as the iP's were different. But shortly after that, I lost RDP into the FreeBSD box. When I booted up the direct monitor and keyboard on the box, there was a mostly black screen with a desktop nav bar at the bottom and a black X mouse cursor.

I rebooted the machine and there was no Desktop and I cannot access via SSH

Seems odd that the Raspberry Pi restoration might mess with the box, but ....

Also, I did get a message on the box that there was another instance of the hosts file name for the box on the network. Not sure how that might have occurred, but possibly from the Raspberry Pi backup?

I disconnected the Pi and still same behavior from the box.

I also noticed there is a nonexistent folder listed in my $PATH, but seems like it was a mistake from a long time ago. "...home/db/bin"
 
dbtrade said:
Where would you start troubleshooting this?
Click to expand...
dbtrade said:
What would you do?
Click to expand...
On the console. Boot it. Look for error messages. Login as root and investigate, see if sshd(8) is configured to start at boot for example. Run service sshd start and see if it starts properly. Check /var/log/messages, again, look for errors and/or clues why things might be failing.

After you did the major version upgrade, did you also reinstall all ports/packages? That's a common mistake new users seem to make.
 
SirDice said:
On the console. Boot it. Look for error messages. Login as root and investigate, see if sshd(8) is configured to start at boot for example. Run service sshd start and see if it starts properly. Check /var/log/messages, again, look for errors and/or clues why things might be failing.

After you did the major version upgrade, did you also reinstall all ports/packages? That's a common mistake new users seem to make.
Click to expand...
I did do the package upgrades before all this started based on reading a thread where I think you were advising the best way to do that.

I ran the "service sshd start" and it advised me to add "sshd_enabled="YES" to my /etc/rc.conf.
It's been a while since RDP wasn't working so I suppose it was possible that I hadn't been using SSH? Seems unlikely... lol thanks for your patience as I know how this must read.

Going to try re-setting up X11VNC which is how I think I was accessing via RDP...

And I am assuming your advice is to just start adding back the stuff that is not working, which might be the desktop next.
And I don't need to be concerned that I was using sudo -1 and su intermixed, although not at the same time

(later)
Looks like when I installed and re-installed llvm13-13.0.1_6 because it reported as missing shared files after I downgraded a bhyve component trying to get my vm's to boot, it removed some Xorg packages I likely need to XFCE4. I found this after following your suggestion to check the log/messages.

I won't mark it solved yet, but I very much appreciate your sanity check on my thinking.
 
SirDice said:
Well, that explains why sshd(8) wasn't running, it wasn't configured to start at boot.
Click to expand...
The problems started after I downgraded a bhyve component while trying to get my vm's to boot, edk2-bhyve-g202308_4 with edk2-bhyve-g202202_1.pkg.
When that made no difference, I re-installed the 202308_4 version.

I ran a sanity check on pkgs, before and after each of those and after I restored the edk2-bhyve, "pkg check -d" and "pkgcheck -B" reported the llvm12, llvm13, & llvm15 were missing shared components. I removed them and re-installed them and the check reported no issues. Although I suspect I missed something else being removed when I removed those packages.

But that is when things went south beyond the vm's not running.

There is a lot more that is wrong, such as cannot get VNC, RDP running automatically despite trying many different ways of doing that. Yet Xrdp shows as running process, I just can't access it via rdp.
I won't go on as it gets to be too much.

At this point, I have paused my efforts and want to ask advice on this.

Would it be feasible to downgrade back to 12.4, examine things thoroughly and then re-upgrade to 13.4? It's been about 8 days since the upgrade to 13.4.
Or am I likely asking for more pain?

I am grateful for any thoughts on the way through this.
 
I am tackling issues one at a time. Would it be better to start a new thread for each one?

I am noticing odd things.
Running as su root from the local console, and then running as su root from a vnc, I get a different command history.

Trying to start x11vnc from ssh as su root, it cannot decode the password file.
Starting x11vnc as su root from the local console, succeeds.

Is this all normal?
 
I have a snapshot of the OS drive when it was 12.4 p8...
It lives as a file on a storage drive.
Although I have never attempted to restore a system drive from backup,
I am wondering if now is the time to do it with what seems to me, bizarre behavior.

Any thoughts?
 
If you’ve got another machine why not try 13.2 on there as a fresh install and see if you can get everything working on there the way you expect?

If you can then you know it’s not issues with 13.2 as such but something that happened during your upgrade.

And if you have a working 13.2 system you can compare what’s different with the broken system versus the working system and see if you can get the broken system up to 13.2.

Your posts each seem to contain five or six possibly unrelated issues so difficult to follow any of it. I don’t think any point starting separate threads for them because they might all be related to the same underlying issue(s).

If su worked for you previously just go back to using it. Yes, user/root history can be different depending on how you log in.

Take a deep breath and start with a clean 13.2 install. Or start with a 12.4 install - set things up the way you want (at least enough of your previous setup to test) and try the upgrade process again but very carefully - see if it works or if you have the same issues as previously.
 
Your posts each seem to contain five or six possibly unrelated issues so difficult to follow any of it. I don’t think any point starting separate threads for them because they might all be related to the same underlying issue(s).
Click to expand...
You are right. It's really hard to discern what is related and what is not.
 
I got this all working, however, I don't have any sure steps to help anyone in a similar situation.
I mucked about with the Xrdp ini and configs. Reset the VNC passwords making sure to remove other password files/locations and all set with root.
Did some related package installs...
Marking this solved though it will be unhelpful to any trying to follow it.
I have no real clue why it all went south, but it was around the time of downgrading firmware needed to Bhyve-vm's to run. I must have messed that up somehow.

273560 – sysutils/edk2: bhyve flavor causes UEFI VMs to fail to boot after update to 202308

bugs.freebsd.org bugs.freebsd.org
 
You must log in or register to reply here.
Back
Top