LXDE Startx hangs

I have a FreeBSD 13.1 installation on a USB device. This has been running lxde without any problem for some time on a ThinkPad X61, although the system has been running with a fan problem for some time which has caused the sytem to overheat and turn off after 15 mins or so.
I have just replaced this system with an 'identical' X61 and am booting from the same USB device. Unfortunately on the 'new' system X hangs almost as soon as it starts and I need to power off to halt the system.

Any ideas on how I can set about debugging this problem?

Would dmesg indicate that it hung once I rebooted?
 
the 'new' system X hangs almost as soon as it starts

Easyest way would be to boot into the TTY console (no direct GUI starter),
then give the full startx command that is in your .shrc or .xinitrc and
watch the output. The error might be in the last line.
 
Tried ssh?

Or virtual console?


Are the BIOS versions of both machines identical?
Good suggestion using ssh from another box.

Not sure how to use a virtual console.

Looking at BIOS shows some differences, although not sure what to make of them. The newer machine has an older BIOS, has a less powerful CPU (1.80GHz as opposed to 2.10GHz) but has more memory (4GB v 2GB). Not sure if this would account for anything.

Since I started monitoring the problem, it hasn't re-occured, but I'll know what to do if it does.
 
You quoted in your post the link to the handbook - exactly that link explains it ;)

I probably used this in the past, but have forgotten about it...
When using Xorg as a graphical console, the combination becomes Ctrl+Alt+F1 to return to a text-based virtual console.

Not sure how to get back to Xorg though.....
 
When you're running this stuff from a USB drive, you should move certain parts of the filesystem to RAM using tmpfs mounts. You should also consider the USB device throughput when looking for performance bottlenecks. Xorg always takes a few seconds for me when it starts which I think is due to CPU usage and shared memory. Try setting kern.sched.preempt_thresh=224 in /etc/sysctl.conf and see if that makes any difference.
 
When you're running this stuff from a USB drive, you should move certain parts of the filesystem to RAM using tmpfs mounts. You should also consider the USB device throughput when looking for performance bottlenecks. Xorg always takes a few seconds for me when it starts which I think is due to CPU usage and shared memory. Try setting kern.sched.preempt_thresh=224 in /etc/sysctl.conf and see if that makes any difference.
A value kern.sched.preempt_thresh=120 is often better for a desktop. A very high value like 224, means that to many processes will try to get a high priority. That is counter productive.
 
I mentioned the virtual consoles as a way to check if you really need to power cycle the machine. Has it really crashed that hard? Or can you switch to a text virtual console and login and safely shutdown the machine? You can also check any logs before rebooting.

I asked about BIOS because I had a Lenovo ThinkPad NOT the same as yours - at some point if I rebooted it failed to start X. Cold starts were fine, just the reboot. So not the same symptoms as yours.

Updated the BIOS and X was fine from cold starts or restarts.

The fact that you have machine A with a newer BIOS that works and machine B with an older BIOS that doesn’t work points towards a BIOS update being something to consider.
 
I mentioned the virtual consoles as a way to check if you really need to power cycle the machine. Has it really crashed that hard? Or can you switch to a text virtual console and login and safely shutdown the machine? You can also check any logs before rebooting.

I asked about BIOS because I had a Lenovo ThinkPad NOT the same as yours - at some point if I rebooted it failed to start X. Cold starts were fine, just the reboot. So not the same symptoms as yours.

Updated the BIOS and X was fine from cold starts or restarts.

The fact that you have machine A with a newer BIOS that works and machine B with an older BIOS that doesn’t work points towards a BIOS update being something to consider.
I will update the BIOS as a matter of course, but as a result of this thread I will have means to track down the cause of the problem, although it does not occur consistently.
 
Back
Top