Ugo Bellavance wrote: > Ugo Bellavance wrote: >> Alexander Dalloz wrote: >>> Ugo Bellavance schrieb: >>> >>>> I found something. >>>> >>>> The script does a -HUP on sendmail, so here is what I see in maillog is: >>>> >>> Why does it do so? >> Because the script's purpose is to sync sendmail configs among many mail >> servers. >> >>>> Jul 26 10:00:02 server sendmail[19016]: restarting /usr/sbin/sendmail >>>> due to signal >>>> Jul 26 10:00:02 server sendmail[19018]: restarting /usr/sbin/sendmail >>>> due to signal >>>> Jul 26 10:00:02 server sm-msp-queue[19032]: restarting >>>> /usr/sbin/sendmail due to signal >>>> Jul 26 10:00:02 server sendmail[25475]: k6QE01aW025475: from=root, >>>> size=0, class=0, nrcpts=0, relay=root at localhost >>>> >>>> >>> That is not really a mail: it has zero size and zero recipient. Could be >>> a script that probes whether the MTA is alife. >> Ok. >> >>>> Jul 26 10:00:02 server sendmail[25579]: starting daemon (8.13.7): >>>> queueing at 00:15:00 >>>> Jul 26 10:00:02 server sendmail[25581]: starting daemon (8.13.7): SMTP >>>> >>>> >>>> It looks like sendmail may not be ready yet... >>>> >>>> >>> Sendmail restarts as to what the log says. >>> >>>> Any ideas? >>>> >>> `mailq -Ac -v' has content? Then there is a problem with mail >>> submission, as cron directly injects the mail. >> /var/spool/clientmqueue is empty >> Total requests: 0 >> > > Any ideas how I could make my script so that I can submit the mail > without problem even if I do a killall -HUP sendmail during the script? > I finally found that the sendmail HUP was the problem. > Regards, > > Ugo