killall -HUP sendmail
in [Sendmail]
|
From: RICCARDO on 10 Feb 2010 03:39 I migrated from Fedora Core 1 + Sendmail 8.12.10 to Debian Sarge + sendmail 8.14.3-5 and I usually send "killall -HUP sendmail" to permit sendmail to read again configuration file (e.g. /etc/mail/access). Now I get this error: sendmail: no process killed Why it doesn't work now for this new o.s. (Debian Sarge) ?!
From: D. Stussy on 10 Feb 2010 15:41 "RICCARDO" <ric.castellani(a)alice.it> wrote in message news:227508f7-9046-482e-8089-356e77fdd883(a)r33g2000yqb.googlegroups.com... > I migrated from Fedora Core 1 + Sendmail 8.12.10 to Debian Sarge + > sendmail 8.14.3-5 and I usually send "killall -HUP sendmail" to permit > sendmail to read again configuration file (e.g. /etc/mail/access). > Now I get this error: > > sendmail: no process killed > > Why it doesn't work now for this new o.s. (Debian Sarge) ?! A running sendmail cannot reload its configuration on the fly. What is done to "reload" is that it execs itself so that it reads the configuration as a new process (reusing the same process number). To call exec(), it needs to have been started by specifying its ENTIRE pathname. Being in the search path and specifying the command name without a path might work depending on the implementation and which variation of exec() in the library is used (i.e. the exec*p() versions). If your sendmail program wasn't started by specifying its complete pathname, this may be the cause of the error.
From: RICCARDO on 11 Feb 2010 05:30 On 10 Feb, 21:41, "D. Stussy" <spam+newsgro...(a)bde-arc.ampr.org> wrote: > "RICCARDO" <ric.castell...(a)alice.it> wrote in message > > news:227508f7-9046-482e-8089-356e77fdd883(a)r33g2000yqb.googlegroups.com... > > > I migrated from Fedora Core 1 + Sendmail 8.12.10 to Debian Sarge + > > sendmail 8.14.3-5 and I usually send "killall -HUP sendmail" to permit > > sendmail to read again configuration file (e.g. /etc/mail/access). > > Now I get this error: > > > sendmail: no process killed > > > Why it doesn't work now for this new o.s. (Debian Sarge) ?! > > A running sendmail cannot reload its configuration on the fly. What is > done to "reload" is that it execs itself so that it reads the configuration > as a new process (reusing the same process number). To call exec(), it > needs to have been started by specifying its ENTIRE pathname. Being in the > search path and specifying the command name without a path might work > depending on the implementation and which variation of exec() in the > library is used (i.e. the exec*p() versions). > > If your sendmail program wasn't started by specifying its complete > pathname, this may be the cause of the error. My scope is to restart process, infact in previous Linux os I'm usually send Hang-up signal to sendmail and I'd like to do same thing. Excuse but in my previuos words there is been misleading.
From: Claus Aßmann on 11 Feb 2010 09:15 RICCARDO wrote: > sendmail 8.14.3-5 and I usually send "killall -HUP sendmail" to permit > sendmail to read again configuration file (e.g. /etc/mail/access). Don't do that. You are terminating all connections. DB maps are reloaded dynamically anyway. If you want sendmail to read the cf file again (or other text files), send a HUP signal to the daemon only, not to any other session oriented process.
|
Pages: 1 Prev: MX records for smart hosts Next: killall -HUP sendmail [Linux/Debian/Sarge] |