Green rectangles appear, system is immediately unresponsive and then reboots

[Palu]

I just completed a new build and think I did comprehensive hardware research for Linux incompatibility issues, but after a silky smooth installation I logged in and got green lines / rectangles after about a minute. It doesn't seem to be based upon particular actions. It happens even if I wait on the login screen. Before it happens I can connect to internet, load the software center, etc. Afterwards the system doesn't respond to mouse or keyboard. Google has plenty of ideas for Windows problems, but nothing that seems to apply to me. I don't know if this is a driver or a bad processor, etc.

OS: Ubuntu 12.10, 64-bit
Processor and Graphics: i5 3570k intel HD 4000
HD: Samsung 840 Pro Series, 256 GB SATA 6GB/s SSD (MZ-7PD256BW)
MB: Gigabyte Mini-ITX (GA-Z77N-WIFI) LGA 1155
RAM: Corsair Vengeance 16GB (2x8GB) DDR3 1600 MHz

Troubleshooting so far:
1. Swapped HDMI in for the DVI I started with before realizing the system was unresponsive too.
2. I thought I should try booting from a USB (for instance, the one I installed Ubuntu from) which I cannot seem to do because the computer doesn't seem to have a bios screen that stays on long enough for me to change boot order. It is instantly at GRUB. I guess that's the speed I get with and SSD... I feel like this would be a good test, so any tips are appreciated. Can I add a USB to the GRUB menu or select it somehow? I don't see an obvious option.

greenrec.jpg

[matt_symes]

Hi

No idea what could be causing that.

The first place to look would be the logs though.

I suspect it's the only way to get a handle on what is happening.

When the system freezes can you get to a console (ALT + F1) and restart lightdm ?

sudo service lightdm restart

Does the magic key combination reboot gracefully ?

alt + sysReq (keep depressed then hit the next 6 keys, one after another, with a 3 second gap between) r e i s u b

I.E. Is it X or the kernel hanging ?

Post

/var/log/xorg.0.log
/var/log/dmesg
/var/log/syslog

Zip then up and add them as attachments in your next post.

Obviously make sure you get the correct version of the logs. i.e. get the xorg log file for when the system was actually playing up.

As for the BIOS screen, start tapping the BIOS key before starting the PC and hammer it hard

Kind regards

[Palu]

I just booted from the USB I installed from and went into the "try Ubuntu" mode. Same results, so I think that rules out the SSD. I tried your key combo too. The first attempt it restarted before I'd typed it all. The second time, nothing seemed to happen. Since the mouse and keyboard don't respond, I cannot get into a console, but I'm looking through my usb now from this computer to see if I can access logs from the USB. If it looks like the logs are empty, I'll put the machine's SSD in an HD dock and find them on there.

[matt_symes]

Hi

If the magic key combination is not working then that would point to the kernel locking up.

One thing i would suggest is to practice the magic key combination when the computer has not locked up.

It can be a little bit difficult to get right and, if you try when the computer has not locked up, you will be able to see if you have got the combination correct.

Since the mouse and keyboard don't respond

What you should see is the computer reboot.

The distinction between the kernel crash and X hanging is quite large and it is important to find out which one is happening.

The keybaord does not have to look like it's responding to do the magic key combination. As long as the kernel has not locked up it should work.

Obviously, as you pointed out, this is not the same when trying to get to the console.

Kind regards

[Palu]

I think I'm doing the magic key correctly because I did it 4 times now pre-crash but post crash it only worked once, so I think that was a fluke of that simply being the arbitrary time at which the system would have reboot anyway.

Just now (pre-crash) I hit CTRL ALT F1 to switch to the console, and I don't have a crash yet. It's been a minute or two longer than the normal time till crashing. I'm looking up how to find the log files and zip them up now. (I'm not very good with terminal, but I'm eager enough.) Will that still help to get the log file pre-crash, or do they get wiped on reboots?

[Palu]

Ok, so... I used ls, cd, zip and ftp to get the file to myself on another computer. See attachments. I included more than just the logs you asked for... mylogs.zip If the attachment doesn't work, it's also on my ftp site: http://www.danwolf.net/mylogs.zip

[matt_symes]

Hi

It does sound like a kernel crash !

I see many of these in kern.log

Mar  2 09:33:34 dan-box kernel: [    0.000000] *BAD*gran_size: 128K     chunk_size: 16M     num_reg: 10      lose cover RAM: -8M
Mar  2 09:33:34 dan-box kernel: [    0.000000] *BAD*gran_size: 128K     chunk_size: 32M     num_reg: 10      lose cover RAM: -8M
Mar  2 09:33:34 dan-box kernel: [    0.000000] *BAD*gran_size: 128K     chunk_size: 64M     num_reg: 10      lose cover RAM: -8M

<snip>

Mar  2 09:33:34 dan-box kernel: [    0.000000] mtrr_cleanup: can not find optimal value
Mar  2 09:33:34 dan-box kernel: [    0.000000] please specify mtrr_gran_size/mtrr_chunk_size

I'll keep looking.

Kind regards

[Palu]

Do you recommend I install all current available updates, or would that potentially make troubleshooting more complicated? I could do this from the command line since it seems stable there, and as a new install, there could be a few.

[matt_symes]

Hi

There is also this in your log file

.980514] ACPI Warning: 0x0000000000000460-0x000000000000047f SystemIO conflicts with Region \PMIO 1 (20120320/utaddress-251)
Mar  2 09:28:14 dan-box kernel: [    6.980518] ACPI: If an ACPI driver is available for this device, you should use it instead of the native driver
Mar  2 09:28:14 dan-box kernel: [    6.980519] lpc_ich: Resource conflict(s) found affecting iTCO_wdt
Mar  2 09:28:14 dan-box kernel: [    6.980521] ACPI Warning: 0x0000000000000428-0x000000000000042f SystemIO conflicts with Region \PMIO 1 (20120320/utaddress-251)
Mar  2 09:28:14 dan-box kernel: [    6.980523] ACPI: If an ACPI driver is available for this device, you should use it instead of the native driver
Mar  2 09:28:14 dan-box kernel: [    6.980525] ACPI Warning: 0x0000000000000500-0x000000000000053f SystemIO conflicts with Region \LED_ 1 (20120320/utaddress-251)
Mar  2 09:28:14 dan-box kernel: [    6.980526] ACPI Warning: 0x0000000000000500-0x000000000000053f SystemIO conflicts with Region \GPIO 2 (20120320/utaddress-251)
Mar  2 09:28:14 dan-box kernel: [    6.980528] ACPI: If an ACPI driver is available for this device, you should use it instead of the native driver
Mar  2 09:28:14 dan-box kernel: [    6.980528] lpc_ich: Resource conflict(s) found affecting gpio_ich

I'll keep looking at the logs.

As an initial test, can you remove everything from your box bar the motherboard, memory (take out all but one stick), graphics card, ssd drive and keyboard. connect to a monitor.

Boot with that. Does it still freeze ?

Kind regards

[matt_symes]

Hi

I don't doubt the machine is totally locking up.

The thing i don't understand about this is there is no kernel panic call stack at all.

It's almost like the hardware itself is locking up (hardware, BIOS, ACPI).

Double check every single BIOS setting you have. Strip the PC down to bare bones and boot it to a LiveCD/USB, incrementally adding hardware.

You said the PC still froze from a LiveCD/USB ? Then you need to do this.

I'm wondering if those mtrr problems are the root cause of the issue.

Kind regards