Re: Possible bug v7.3-6.el5: "Can't exec "sendmail"...

----- "Michael Tautschnig" <[email protected]> wrote:
> > My apologies if this is not the appropriate place to post this
> inquiry in advance.
> > 
> > The system in question is CentOS 5 based, which were all recently
> upgraded from v7.3 to v7.3-6.el5.  Prior to this the configuration
> file was set to email the generated reports to a monitoring "catch
> all" account and was working fine.  After the upgrade, the email
> reports are not coming through.  The email service is Postfix.  A
> quick and simple email test via mutt shows that email is going to the
> "catch all" account just fine from the system.
> > 
> > Running the cron job from /etc/cron.daily generates the following
> error message:
> > 
> > Can't exec "sendmail": No such file or directory at ./0logwatch line
> 1010, <TESTFILE> line 3.
> > Can't execute sendmail -t: No such file or directory
> > 
> > Prior configurations on other systems that we monitor show this line
> in the /usr/share/logwatch/default.conf/logwatch.conf file with this
> same line, so no changes to the configuration file can be seen there.
> > 
> > Thoughts and comments are welcomed.  TIA.
> > 
> 
> I've got no idea about CentOS specifics, but you definitely need to
> set the
> "mailer" variable in logwatch.conf to an appropriate value, wherever
> sendmail is
> to be found on your system. In Debian, e.g., we've got
> /usr/sbin/sendmail.
> 

The logwatch.conf file has

mailer = "sendmail -t" 

As noted above, it was working, now it is not which seems to be the centered area of the issue.  In Debian/Ubuntu based machines I have seen this as 

mailer = /sbin/mail

as well which I am guessing is the modification between distros here.  As noted in the logwatch.conf comments above, the author is/was going for a "one does" all for mail programs for those wishing to have the results returned as such.  I am guessing something got tweeked here perhaps since the upgrade from v7.3 to a patched version seems to be the issue.  Worst case, I might see about rolling the systems back for this and seeing if this resolves things or not.

--- Crawford