[CentOS] slowness in sendmail - 60 second timeout

Simon Billis simon at houxou.com
Mon Feb 8 17:08:00 UTC 2010

Jerry Geis sent a missive on 2010-02-08:

>> /Sorry to be more specific - make sure that devcentos5x64.msgnet.com
>> is resolvable.
> from pagestation.com I can "ping devcentos5x64.msgnet.com" - this
> entry is in the /etc/hosts file on that machine.
> There is no official MX record for devcentos5x64.msgnet.com there is
> for pagestation.com.

I'm not clear that there is a dns/host entry for devcentos5x64.msgnet.com on
devcentos5x64.msgnet.com. I would make sure that there is an entry for this
host on devcentos5x64.msgnet.com. If there is then apologies for not getting
it :-).

> on pagestation.com I have the /etc/hosts entry and /etc/mail/access
> entry for devcentos5x64.msgnet.com Perhaps its an MX lookup timeout
> value.
>  I am looking for a timeout setting or a way to place an entry in a
> config file that says "I know this server or machine", trust it and
> dont delay. Is there such a setting?

I think that sendmail is checking the "MAIL" command  - there are no
timeouts that are specific to this rule afaik and sendmail would use a
timeout of 1 second (which is doubled to a max of 4 mins) if this was an
issue with bogus commands or repeated commands. From your description, it
"feels" as if there is a problem with the name resolution of
devcentos5x64.msgnet.com. You may wish to enable FEATURE(delay_checks)dnl in
the sendmail.mc and rebuild the sendmail.cf file and see if it helps. If
this makes a difference then I would look in the MAIL command checks.


More information about the CentOS mailing list