2

For no apparent reason, my machine is rebooting.

last reboot brings lines like this: 

reboot   system boot  3.13.0-24-generi Tue Dec  9 18:57 - 19:09  (00:11)    
reboot   system boot  3.13.0-24-generi Tue Dec  9 12:59 - 19:09  (06:09)    
reboot   system boot  3.13.0-24-generi Tue Dec  9 11:07 - 11:35  (00:28)    
reboot   system boot  3.13.0-24-generi Mon Dec  8 13:52 - 01:20  (11:27)

With the pseudo-user reboot.

I need to know:

  • who is calling/activating this pseudo-user reboot?
  • why(how to track it)?
  • and how to fix/prevent that? (I may have to answer this one myself after I discover what is happening)

I am still researching, but my machine may suddenly reboot, so I have to post now!

related: Why does a grep reboot the machine?, about IO problems
related: about full memory

After two weeks my machine still reboots for no reason. I wonder if anyone has any idea of what "things" should not be accessed? I mean: are there places in /proc or /dev where we should avoid performing too many reads or writes? Or some commands that we should avoid? Do xdotool or wmctrl have some specific commands that should be avoided like trying to command a closed window id or something like that? any idea?

Narrowing it Down
My machine seems to not reboot if I keep some scripts stopped.

I am trying to narrow down what can cause this; I think a script I have may be the source; it, along with others, runs commands like these:

windowId=`xdotool getactivewindow`
xdotool getwindowname $windowId
xdotool search --sync "^Desktop$"
wmctrl -d
wmctrl -i -r $windowId -b toggle,maximized_vert,maximized_horz
xwininfo -wm -id $lnWindowId

Not exactly in this order, but basically they are all related to X window manipulations.
And, they happen very often, like 4 times each second;
And, they may happen on an unavailable window (that was just closed), or on some special windows like conky clock one, or in any non standard "window" (that is not a window like the unity hud).

It seems X windows manipulation can cause reboot then?

Community
  • 1
Aquarius Power
  • 4,289
  • It might be worth posting excerpts from /var/log/messages and your bash history from the time that the reboots were happening. Is this a desktop or a server? – Chris Davidson Dec 09 '14 at 21:21
  • It is a desktop, but I have like 11 bash scripts running simultaneously on misc tasks, but they are like that for about a year or so. I will check that messages. – Aquarius Power Dec 09 '14 at 21:29
  • Nobody calls or activates the user. It's just how a reboot is logged in the records that last uses, no matter what triggered the reboot. – muru Dec 09 '14 at 21:46
  • @Dayvo the reboot happened around 18:33, may be 18:32:30 (I got stunned), on log there was this: "Dec 9 18:30:39 hostid wpa_supplicant[1226]: wlan0: CTRL-EVENT-SCAN-STARTED", and this "Dec 9 18:31:35 hostid postfix/sendmail[3555097]: fatal: open /etc/postfix/main.cf: No such file or directory", but these last ones happen very often and they were much before the reboot time. – Aquarius Power Dec 09 '14 at 21:48
  • @muru, any tip on how to catch what triggered the reboot? – Aquarius Power Dec 09 '14 at 21:49
  • Can you grep for shutdown in the various log files in /var/log as a first step? The exact files varies based on distribution (I don't have a messages, for example). – muru Dec 09 '14 at 21:54
  • @muru, here is /var/log/syslog also. But anyway, at /var/log, this grep shutdown * -irn 2>/dev/null only showed wtmp (but that is what we read at last) – Aquarius Power Dec 09 '14 at 22:10

0 Answers0