I can not help really you with this trouble, but I experience the same
issue with my daily-log-rotate.
I was told this would not be due to log-rotate, but to some other daemon
sending a sig-kill. As I did search for it now since three days not
finding anything, I came to the conclusion this can not be true.

The issue can be temporally solved by adding:
 save-restart)
    DPID=$(ps -ef|grep sogo|grep -v grep)
    if [ "${DPID}" = "" ]; then
     # Ensure directory's existence and permissions
         install -o sogo -g adm -m 755 -d /var/run/$NAME
         install -o sogo -g adm -m 750 -d /var/spool/$NAME
         install -o sogo -g adm -m 750 -d /var/log/$NAME
         start-stop-daemon -c $USER --quiet --start --pidfile $PIDFILE
--exec $DAEMON -- $DAEMON_OPTS
     log_progress_msg "Had to restart sogo"
         log_end_msg 0
    else
     log_progress_msg "Sogo still running, doing nothing"
     log_end_msg 0
        fi
    ;;

to /etc/init.d/sogo and adding  "55  7 *   *   * /etc/init.d/sogo
save-restart" via crontab -e to my cronjobs (in my case sogo dies around
7:50 each morning, you might want to adapt the timing to your log-rotate
timing).
If you find out more, I would be pleased to know about your findings ...

cheers,
Michael

On 14.03.2011 12:40, Alessio Fattorini wrote:
> Can you help me about this trouble? During weekly logrotate, sogo stop
> and doesn't restart.
>
> This is old log
>
> # tail /var/log/sogo.log.1
> Mar 03 04:02:13 sogod: SOGo watchdog [2694]: <0x0x84ea618[WOWatchDog]>
> Terminating with signal 15
> Mar 03 04:02:18 sogod: SOGo watchdog [2694]:
> <0x0x8506188[WOWatchDogChild]> sending terminate signal to pid 2695
> Mar 03 04:02:29 sogod: SOGo watchdog [2694]:
> <0x0x8506188[WOWatchDogChild]> child 2695 exited
> Mar 03 04:02:29 sogod: SOGo watchdog [2694]: <0x0x84ea618[WOWatchDog]>
> all children exited. We now terminate.
>
>
> New sogo.log is empty :-|
>
> This is logrotate script:
>
> #cat /etc/logrotate.d/sogo
> /var/log/sogo/*.log {
>         missingok
>         rotate 5
>         notifempty
>         create 0664 sogo sogo
>         postrotate
>                 /sbin/service sogod restart > /dev/null 2>&1
>         endscript
> }
>
>
> No errors in messages.
>
> I can't reproduce that fail, if i execute logrotate manually works.
> If i execute restart manually, it works.
>
> Service is a link to /etc/rc.d/init.d/sogod
>
> This is verbose mode of logrotate. It works:
>
> # logrotate -fv /etc/logrotate.conf
>
> rotating pattern: /var/log/sogo/*.log  forced from command line (5
> rotations)
> empty log files are not rotated, old logs are removed
> considering log /var/log/sogo/sogo.log
>   log needs rotating
> rotating log /var/log/sogo/sogo.log, log->rotateCount is 5
> renaming /var/log/sogo/sogo.log.5 to /var/log/sogo/sogo.log.6
> (rotatecount 5, logstart 1, i 5),
> renaming /var/log/sogo/sogo.log.4 to /var/log/sogo/sogo.log.5
> (rotatecount 5, logstart 1, i 4),
> renaming /var/log/sogo/sogo.log.3 to /var/log/sogo/sogo.log.4
> (rotatecount 5, logstart 1, i 3),
> renaming /var/log/sogo/sogo.log.2 to /var/log/sogo/sogo.log.3
> (rotatecount 5, logstart 1, i 2),
> renaming /var/log/sogo/sogo.log.1 to /var/log/sogo/sogo.log.2
> (rotatecount 5, logstart 1, i 1),
> renaming /var/log/sogo/sogo.log.0 to /var/log/sogo/sogo.log.1
> (rotatecount 5, logstart 1, i 0),
> old log /var/log/sogo/sogo.log.0 does not exist
> renaming /var/log/sogo/sogo.log to /var/log/sogo/sogo.log.1
> creating new log mode = 0664 uid = 3000 gid = 3000
> running postrotate script
> removing old log /var/log/sogo/sogo.log.6
>
>
>


Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to