Every day at 06:25, my Sheevaplug restarts. This is the line from syslog:
Feb 21 06:25:14 Sheeva_1 syslogd 1.5.0#5ubuntu3: restart.
Isn't that syslogd
restarting? As part of log rotation?
The logs get renamed, then syslgod is sent a HUP to re-open all of its log files.
What makes you think the Plug is restarting. What does uptime report for the system uptime?
Good point, I noticed that the checksum of my two attached USB-disks changes at exactly that moment. That is, cron initiates a comparison of "old" and "new" checksum every hour (hh:00). If the checksums are the same, the disks are spun down. At 06:00 the command is "spin down". At 07:00 the log says "do nothing", meaning the checksums differ.
The conclusion is that "someone" has done "something" on the disks between 06:00 and 07:00. The only thing I find in syslog is this "Feb 21 06:25:14 Sheeva_1 syslogd 1.5.0#5ubuntu3: restart". But cron.daily contains a fair number of scripts that are all executed at 06:25, and perhaps not all of the actions are shown in syslog.
Grateful for any suggestions, I will continue setting up logs and try to find out if really something is happening on the disks. I don't want the disks to spin up every night for no reason.