On Mon, May 12, 2008 at 12:40 AM, Shlomo Solomon <[EMAIL PROTECTED]> wrote:
> On Saturday 10 May 2008, Amos Shapira wrote: > > It's a long shot but maybe the script running exactly two minutes into > > the daily cron jobs causes some interaction that breaks things? > > How about disabling this script for this time of the day and see if > > something is different? > > Another option - try strace'ing it on its 2:02 run. > > Maybe I wasn't clear enough, but 2:02 is always the time of the last log > entry. I don't know exactly what time the crash is or exactly which cron > job I see. > In any case, I've added 2 lines to each daily cron to write to a temp file > at > the satrt and end of each job. I'll check this file after the next crash to > see which job started but didn't end. I know it's a kludge, but that's the > best I could do for now. Consider adding a call to sync(8) after each timestamp file update to (hopefully) make sure that the file update actually happens on the disk before the crash. --Amos
