Every so often, my laptop will crash. The screen will freeze, including the cursor, the lock lights (caps lock and the like) will turn off if they were on, though my battery, power, and wifi lights still work. I'm not sure if it's a kernel panic, because sometimes it will just freeze, then the same sound it makes when I turn it off by the power button, and then a reboot. Rarely, I see a screen of text (not like the booting text) and then it reboots the same way. I think I saw 'kernel' and 'panic' on different lines. But I can't be sure. There was one time when it did the same for a minute, but then I could move the mouse while none of the keys would work. Then it crashed and rebooted.
I'm trying to figure out what's causing it.
Originally, I thought it was a heat issue because it stopped happening when I put my laptop on a cooling fan on a hard surface, and it only was happening when I visited a resource-intensive website. Then I thought it was a memory issue because heat was taken care of, but memory was overflowing into the swap partition. But now it's been crashing even when memory use is low and heat is not setting off alarms in psensor(I have alarms set to go off at 60C for any component, and I'm pretty sure it hasn't gotten to 80C since I stopped placing my laptop on my lap).
I thought my cousin had installed Debian 8.2(Jessie), but the
cat /etc/debian_version returns "stretch/sid". I'm using XFCE, and it's on a Dell Inspiron 1520, which originally had a WindowsXP os(we wiped the hard drive before installing Debian), and has 2Gb RAM. The kernel version is 3.16.0-4-amd64; however, I believe my question is about Linux in general, not the specific version. And my cousin did run memtest86 from a dedicated disk, after wiping the hard drive, before we installed any OS.
I don't what I need to check or how to check it, how to determine what's a kernel problem, etc. Here's what I've heard I need to look at(scite only there for opening text files):
- scite /var/log/kern.log
- scite /etc/rsyslog.conf
- scite /etc/rsyslog.d/
- scite /etc/rsyslog.d/50-default.conf
- scite /var/log/syslog
- sudo /usr/sbin/mcelog > mcelog.out
- scite /var/log/dmesg
- scite /var/log/syslog.1
- scite /var/log/messages.1
Several of these are blank or the equivalent thereof, such as var/log/dmesg, with it's measly one-line of (Nothing has been logged yet). And I can't find a dmesg.1 in /var/log
I wasn't able to isolate anything in the logs until recently, because I'd always just let my computer restart. This most recent time, I forced it to shutdown instead of riding through the panic. And I've left it down for a few days. The syslog contains nothing from before today. In the kern.log, the last thing from the day it last shut down was this:
Oct 17 22:51:27 darty kernel: [ 299.816045] mce: [Hardware Error]: Machine check events logged
Everything before that is very repetitive, and is identical to things 5 minutes prior to the error. (It's been shutting down within five minutes in some instances.)
sudo /usr/sbin/mcelog > mcelog.out returns nothing, though the file contains a bunch of gobbeldygook; I'm assuming it's machine code or something.
/var/log/syslog.1 contains some... interesting stuff... A lot of it is just like the kern.log, lots of stuff about nouveau Iceweasel... other repetative stuff. At Oct 17 22:13, there's a big wall of NUL characters all on one line, but that's about 40 minutes before the last crash. I think I crashed a few times on the way though... Oof... After literally the exact same line as the one above, but in the syslog.1 file, is another big wall of NUL, followed by
Oct 19 17:39:18 darty rsyslogd: [origin software="rsyslogd" swVersion="8.4.2" x-pid="1547" x-info="http://www.rsyslog.com"] start
Still doesn't tell me anything, other than that I'm definitely not sure what version of Debian I'm on (The boot usb is all 8.2, I swear. So why this is 8.4.2, and then it tells me I'm on stretch/sid, ugh...).
The same line, preceded by fewer NUL characters, appears in the /var/log/messages.1 file. Before that are about 45 lines in 27 seconds of the following, except with a seconds-difference time stamp and Source ID ranging from 77 to 230, non-consecutively but successively.
Oct 17 22:50:22 darty org.freedesktop.Tracker1: (tracker-store:2019): GLib-CRITICAL **: Source ID 77 was not found when attempting to remove it
But those are also located in places that don't include a crash.
So ... yeah. I'm at a complete loss as to how to determine what the actual error is. I can almost reliably reproduce it by doing certain things on the website CodinGame.com (I know the age-old wisdom is "If it hurts, don't do it." Well, the whole reason I have a computer is so I can do things like CodinGame. The analogy from this to medical would be "Don't stand up or walk around." I can't follow it indefinitely.) Again, it is not heat nor memory. I've got sensors on both of those with widgets(applets?) on the panel. But how do I figure out where the error lies? Can I set something up to capture the crash, then induce it by playing with CodinGame?
Yes, I have read "How to determine why my computer crashed?" and about 5~10 questions related by wording on askubuntu. And How do I figure out what's freezing up my machine?. The answer given for the latter doesn't seem to be my problem because my cousin just installed this system less than a month ago, and I haven't used it for much besides light text editing and file transfer and somewhat heavy (not MMORPG-heavy, tiny-flash-applet-and-HTML5-heavy) browser usage. The answers to the former seem just like the rest of the stuff I've found before, stopping short of the solution.
TL;DR: How does a person go about reading the log files to find an error that crashes her computer, how does she find the error if it isn't in the log files, and what can be done to fix it?