On Sun, August 19, 2007 12:19 pm, James G. Sack (jim) wrote:
> Gregory K. Ruiz-Ade wrote:
>> On Aug 19, 2007, at 8:53 AM, Lan Barnes wrote:
>>
>>> OK, I replaced sendmail w/ postfix because sendmail was taking minutes
>>> to
>>> load on boot. But now when I restart postfix it takes just as long,
>>> indicating that the fault is in some basic configuration, like maybe
>>> the
>>> dreaded reverse lookup DNS issue (I'm just waving my hands  here -- I
>>> have
>>> no idea what I just said).
>>
>> Well, make sure that myhostname and mydomainname are set to what you
>> want them to be (instead of relying upon postfix to figure it out, which
>> can be problematic on a mobile system that might not always have an IP
>> address.
>
> Confirming (my recollection) that MTAs spend a lot of time looking for a
> fully qualified name upon startup. In old days, sendmail would hang for
> 5 minutes (or more?).
>
> The FQ part is perhaps surprising. Be sure you _have_ a fullname. Try
>  hostname
>  hostname -s
>  hostname -f
>
> Perhaps GKRA can correct me but I seem to remember that a domain name
> without a suffix also causes a problem.
>
> Note that the domain doesn't need to be public or real -- it just has to
> be syntactically correct. I use sack-net.pvt. So I get
>  hostname -f
>  jgs.sack-net.pvt
>
>>
>> Also, RH-based systems tend to put the hostname in /etc/hosts as an
>> alias for localhost.  If it's not there you might want to try putting it
>> there.
>>
>> Also, check your syslog, and see if anything is being reported there
>> during postfix startup.
>>
>> If you stop postfix once your system is running, and start it manually,
>> do you get the same delay?
>
> Regards,
> ..jim
>
>

Gus has implied that I don't need it at all. Is this true?

-- 
Lan Barnes

SCM Analyst              Linux Guy
Tcl/Tk Enthusiast        Biodiesel Brewer


-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list

Reply via email to