I'd look at limiting your startup services to see if any of those are causing problems.
I'd start with anacron, at and cron (you can do the fist two from the services settings application under system administration, but I think cron may need to be stopped manually post boot with "sudo /etc/init.d/cron stop". But it may be safer to use this method for everyting you want to stop, since this way, a reboot will reset everything for you. If they don't help, I'd try acpid and apmd, the system loggers, powernowd. Outside of that, there are a number of other services in init.d you can stop per session. Are you sure you don't have any power management setup? What are your screensaver settings? |
billymayday--
Thanks for helping me. OK, one thing I tried this evening: I rebooted and let it sit at the grub menu for about 3 hours. Then I ctrl-alt-del and changed the bios clock to Friday 11:45 pm or so. Then I let it boot up, changed the system clock to about the same date and time, and took it off the automatic get the time from the internet thing. I have been running now for about an hour and a half (the system clock shows Saturday at 1:34 am--actual time is Monday, 11:34 pm), and there have been no auto reboots. I'll let it go overnight and see. Will report back on what this does.... Here is anacrontab: Code:
doug@doug2:~$ cat /etc/anacrontab Code:
[Sun Aug 10 00:33:00 2008] logfile reopened No I am not sure I have no power management set up. If I go to System > Preferences > Power Management, I have three tabs: Under General what shows is Ask me on power button pressed; suspend when suspend button pressed; Never display an icon; Use sound to notify of error. Under UPS Power I have put computer to sleep never; UPS power low do nothing; ups power critically low shutdown; put display to sleep when 23 minutes inactive. Under AC Power: put computer to sleep never; put display to sleep when inactive 40 minutes. That's everything under power management I see--nothing there sounds suspect to me. APCUPSD Monitor reports these events over this time period: Code:
Mon Aug 11 18:39:31 EDT 2008 apcupsd shutdown succeeded doug@doug2:~$ apm No APM support in kernel So I suspect no apmd is installed--but I do have man files for apmd and apm, suggesting they are installed. Ahh--synaptic shows apmd is installed. In short, I don't know what you mean by "If they don't help, I'd try acpid and apmd, the system loggers, powernowd." If this resetting of the clock produces no ill effects, I will then try your method of shutting down cron, and so forth. But I should probably test only one thing at a time, yes? Is it possible this is hardware? |
I missed the Saturday connection earlier. Suggests very strongly to me something in cron.weekly or, perhaps more likely, a gnome process. I wouldn't think hardware, but who knows.
If you can get it to reboot with you clock deception, try booting to the command line (ie kill gnome) and see if it reboots then. See http://www.linuxquestions.org/questi...d-of-x-549292/ for example |
billymayday--
A report: this morning all seemed well. There were no reboots. So I reset the system clock in Gnome and then rebooted and reset the clock in bios. In other words, the experiment gave inconclusive results. (Remember, last time I had this problem was about 4 months ago, and then a week ago.) What is in cron.weekly is shown in post 13 in this thread. Shall I post the contents of any of these as being a likely culprit? Or perhaps simply move them someplace else one by one? I think I did run top when all this was going on to see what might have been triggering the reboot, and I saw nothing. But my recollection might not be accurate, either. Something for next time this all happens. What should I investigate next? Anything particular in gnome? (I did have a problem last week with losing my evolution addressbook, and had to rename the directory ~.gconf/apps/evolution/addressbook. Does that suggest any clues? That was almost a week before the auto reboots started.) Another piece of info: this problem existed on this machine when I had Ubuntu 7.10 on it, and now after the upgrade to 8.04.01. Thanks, billymayday! |
Hi all--
Just a brief update. The machine continues to run well, so there is no telling what really was wrong. At least I learned a bunch of things to check if and when it happens again! On a side note: a WinXP box in the office started shutting down (not rebooting) unexpectedly about 2 or 3 weeks ago, and that continued for a couple of days. It did not do it every hour, just once or twice a day for 4 or 5 days in a row. Strange. Thanks to everyone for all your help! |
It was baaaaaaack!
Yesterday into today the problem returned, as this output shows: Code:
doug@doug2:~$ last 1. It is not exactly every 60 minutes, but about 60 1/3 minutes--note the progression every few hours. 2. After about a day, it goes away on its own--if I can put up with it. 3. The entry at 13:13 was when I rebooted the machine--notice that the next several automatic reboots were at 13 after the hour. This suggests to me it is not a power problem, nor an UPS problem, but something with the computer itself, and probably the software, since it responded to something I did. The last time this happened was August 9, so there does not seem to be anything with the dates here--August 9 then October 1. I post this more to keep a log of it than to ask help. Since the problem seems to have resolved itself (again) after a day, trouble-shooting would be a hit or miss thing. But is this a clue? It was started at the same time as the last reboot. Code:
doug@doug2:~$ ls -alt /var/log/gdm Thanks, folks! |
Again in November
Hi--
Just another report, and this is a good place to keep a record of what is going on. I note the last time was the first of the month or thereabouts, too: Code:
doug@doug2:~$ last reboot We shall see what we shall see. |
DO you have anything in /var/spool/cron or similar directory (this should be where user crontabs are kept)?
If so, post # cat /var/spool/cron/* or at least look for monthly jobs. What's in /etc/cron.monthly? |
billymayday--
Thanks for continuing to watch this and jumping in to help. This is a really puzzler, huh? Code:
doug@doug2:~$ cat /var/spool/cron/* Thanks, billymayday! |
One of the problems is wtmp doesn't go back that far. Clearly the first (?) reboot was at 7:21, and it's rebooted every hour is my guess, which has pushed the minutes out to 23 not 21. I guess it's possible that it started rebooting earlier but the log was rotated around this time.
Can you look in /var/log and see if you have wtmp.1 or similar (I'm assuming here that /var/log/wtmp exists and is the relevant log). If it does, try last -f /var/log/wtmp.1 reboot or similar for relevant file to try and determine the first reboot of the cycle. If you've found the first at 7:21, I'd be thinking log rotations perhaps. What's in /etc/cron.monthly? |
billymayday--
The first one was the day before at 16:22, so perhaps you are on to something with logrotate: Code:
doug@doug2:~$ sudo last -f /var/log/wtmp.1 reboot Code:
doug@doug2:~$ ls -alh /etc/cron.monthly/ |
Clues, clues?
Hi--
It has continued today a second day--in the past it has stopped after a day or so. If my memory is accurate. Here's what happens, in case it is a clue: I will be sitting typing away and then I hear the computer fan start up. I can save a document--I appear to have about 3 to 5 seconds, and then it goes into a reboot. Here is the latest string of reboots: Code:
doug@doug2:~$ last reboot Thanks! |
Clues?
Hi--
Is this a clue? I'm not sure what it is saying: Code:
doug@doug2:~$ less /var/log/cron.log |grep boot |
No, but since you've got a pretty good idea when it will happen next, try capturing all the processes that are running after the fan fires up (ps aux > ps.file) or something. Or better yet, a minute or so before you expect it to go down, run a script that writes that output to a file every second or so - something like
Code:
#!/bin/sh |
billymayday--
That's worth a shot! Will let you know what happens. Thanks, billymayday! |
All times are GMT -5. The time now is 02:21 AM. |