Conversation
Notices
-
still breaking my head over how to find out what (if anything) happened to my cron jobs; finally http://bencane.com/2011/11/02/did-my-cronjob-run/ gave me a clue: Hal (ASUSTOR NAS) has /etc/rsyslog.conf which has the line '#cron.* -/var/log/cron' - and this: '*.info;mail.none;authpriv.none;cron.none -/var/log/messages' aha! never mind enabling syslog when cron logging is doubly disabled; there is also the fact that 'crontab' is a busybox command here so options may be (even) more limited... off to experiment more...
-
@mk the problem with /etc/rsyslog.conf is of course that it's under the system root, thus loaded from flash memory at boot time - so any changes will be completely wiped out on reboot :(
-
@mk yay! I FINALLY managed to get (busybox...) crond to log to a log file; though setting the log level via '-l N' does not seem to be supported (but there may be a workaround for that via '-d N'). So now I know that a) the updatedb command (to update locate database) is actually running and b) that it is unsuccessful because the output file is locked. #real #progress
-
@mk ...and slowly, slowly, am I getting this nice #NAS box called Hal a bit more under control. Lots of riddles and #frustrations today, but also #learning, #discoveries, and #progress. In all, a good day (in spite of seemingly having a cold!). And more things to figure out #tomorrow. Like: *which* changes that I made *actually* made cron start writing to a log file? and: how do I replace a @reboot crontab rule because BusyBox doesn't support it? There are probably 2, maybe multiple ways... #hammock, for now!
-
@mk This morning, with more entries in the cron log (and even the system log) I'm starting to notice some odd patterns... 1) after ONE manual ntpdate update 23:24 , I am seeing THREE updates about an hour later (all at the same time 00:30) in syslog - but only TWO entries in the cron log - after that, no more entries. 2) Meanwhile, at 23:02 (BEFORE the automated ntpdate) an EXTRA crond was started, while the old one apparently continued running. 3) From then on, the automated updatedb is successful for the NEW crond while the old crond (not surprisingly) encounters a locked mlocate.db file. Both crond's do ONE ntpd update at 00:30 - it's not clear where the third one (recorded in syslog) at the same time came fery hour one successful and one failed updatedb - but nothing else (no ntpdate) is recorded in the log... #nas #cron #confuzzled
-
@mk what's also odd is that this second cron can not have been started with the same command-line parameters I gave the first one - but it's logging to the cron log anyway. ??? #evenmoreconfuzzled
-
@mk ...and in the process list on ADM I see THREE crond's running (that would explain the three ntpdate updates, but not why they stopped again after that)
-
@mk as I now found out, when I STOP crond, it gets automatically restarted moments later. WTH is doing that? (Anyway, no need for a 'restart' option in the (boot) startup script...). What is worse is that sometimes it's started even when there is another crond (or 2) already active. Does 'crontab_check' have anything to do with it??
-
The messages log gave a clue: a process called 'watchmand' "relaunches" crond - no idea why because so far it's been doing this twice while crond was already running...
-
@mk and I cannot find anything anywhere about 'crontab_check' (except that it's apparently NOT part of BusyBox). Asking it for its help only produces an error message that it's already running.
-
-
-
-
-
