How to figure out why a computer rebooted?
I'm disturbed that I've received a Reboot logcheck report from a computer to which I don't have physical access right now. I logged in by ssh and looked at /var/log/syslog, which shows routine stuff, then a 3-minute gap, then the restart log entries. Nov 16 08:02:01 argon /USR/SBIN/CRON[16445]: (logcheck) CMD ( if [ -x /usr/sbin/logcheck ]; then nice -n10 /usr/sbin/logcheck; fi) Nov 16 08:02:01 argon /USR/SBIN/CRON[16446]: (root) CMD (/root/bin/minute) Nov 16 08:03:01 argon /USR/SBIN/CRON[17248]: (root) CMD (/root/bin/minute) Nov 16 08:04:01 argon /USR/SBIN/CRON[17251]: (root) CMD (/root/bin/minute) Nov 16 08:05:01 argon /USR/SBIN/CRON[17255]: (root) CMD (/root/bin/minute) Nov 16 08:06:01 argon /USR/SBIN/CRON[17259]: (root) CMD (/root/bin/minute) Nov 16 08:09:15 argon syslogd 1.4.1#17: restart. Nov 16 08:09:16 argon kernel: klogd 1.4.1#17, log source = /proc/kmsg started. Nov 16 08:09:16 argon kernel: Inspecting /boot/System.map-2.4.27-2-386 Nov 16 08:09:16 argon kernel: Loaded 18328 symbols from /boot/System.map-2.4.27-2-386. Nov 16 08:09:16 argon kernel: Symbols match kernel version 2.4.27. Nov 16 08:09:16 argon kernel: Loaded 787 symbols from 36 modules. A few days ago I deliberately shut the machine down for a while (with physical access) and produced the following bit of syslog. Nov 12 12:56:01 argon /USR/SBIN/CRON[21427]: (root) CMD (/root/bin/minute) Nov 12 12:57:01 argon /USR/SBIN/CRON[21430]: (root) CMD (/root/bin/minute) Nov 12 12:58:01 argon /USR/SBIN/CRON[21433]: (root) CMD (/root/bin/minute) Nov 12 12:58:19 argon gdm[3158]: Master halting... Nov 12 12:58:19 argon shutdown[3158]: shutting down for system halt Nov 12 12:58:19 argon init: Switching to runlevel: 0 Nov 12 12:58:24 argon xfs[2254]: terminating Nov 12 12:58:25 argon xfs[2257]: terminating Nov 12 12:58:27 argon chronyd[30148]: chronyd exiting on signal Nov 12 12:58:27 argon rpc.statd[2457]: Caught signal 15, un-registering and exiting. Nov 12 12:58:27 argon exiting on signal 15 Nov 12 13:16:21 argon syslogd 1.4.1#17: restart. Nov 12 13:16:22 argon kernel: klogd 1.4.1#17, log source = /proc/kmsg started. Nov 12 13:16:22 argon kernel: Inspecting /boot/System.map-2.4.27-2-386 Nov 12 13:16:22 argon kernel: Loaded 18328 symbols from /boot/System.map-2.4.27-2-386. Nov 12 13:16:22 argon kernel: Symbols match kernel version 2.4.27. Nov 12 13:16:22 argon kernel: Loaded 787 symbols from 36 modules. Obviously if I get home and the digital clocks are flashing, I'll know there was a brief power cut. But is there any way to determine from the computer's own evidence why it restarted? Thanks, Adam -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How to figure out why a computer rebooted?
Obviously if I get home and the digital clocks are flashing, I'll know there was a brief power cut. But is there any way to determine from the computer's own evidence why it restarted? Look at this part of your second log, where you shut it down yourself: Nov 12 12:58:19 argon gdm[3158]: Master halting... Nov 12 12:58:19 argon shutdown[3158]: shutting down for system halt Nov 12 12:58:19 argon init: Switching to runlevel: 0 Nov 12 12:58:24 argon xfs[2254]: terminating Nov 12 12:58:25 argon xfs[2257]: terminating Nov 12 12:58:27 argon chronyd[30148]: chronyd exiting on signal Nov 12 12:58:27 argon rpc.statd[2457]: Caught signal 15, un-registering and exiting. Nov 12 12:58:27 argon exiting on signal 15 This is stuff put out by various processes as you go to run level 0 (shutdown), stuff *not* present in the log snippet you gave from the other machine. So it didn't go down through a normal system procedure -- it died abruptly, and then rebooted. Typically, as you note, this is because of a power issue (and depending on whether your machine has e.g. a good UPS with a conditioner, etc, it need not be a failure; I've had bad power quality briefly where the digital clocks in the room didn't go to flashing 12, but the computer power supply allowed the motherboard power to drop enough for the motherboard to quit, then reboot shortly after). I've also seen other times when this has happened because of hardware problems. I've never seen it because of an OS or other software issue -- for that, I'm more likely to get hard lockups or, I guess, PANICs -- but maybe someone will chime in with a counter- example. But whatever the source, it died instantaneously, and in such circumstances the system doesn't get much of an opportunity to say I'm going down because of X. So the only suggestion I'd have would be to look at the logs for any sign of growing hardware issues, which could be located (in the logs) nowhere near the actual shutdown. And check what you're doing about your power quality. -c -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]