Date: Fri, 15 Feb 2008 16:49:11 +0000 From: Anton Shterenlikht <mexas@bristol.ac.uk> To: Bill Moran <wmoran@potentialtech.com> Cc: freebsd-questions@freebsd.org Subject: SOLVED: Re: mailq takes 25 sec to respond Message-ID: <20080215164911.GA45654@mech-aslap33.men.bris.ac.uk> In-Reply-To: <20080215113334.71e3380e.wmoran@potentialtech.com> References: <20080215162713.GA45570@mech-aslap33.men.bris.ac.uk> <20080215113334.71e3380e.wmoran@potentialtech.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Feb 15, 2008 at 11:33:34AM -0500, Bill Moran wrote: > In response to Anton Shterenlikht <mexas@bristol.ac.uk>: > > > On my ds10l alpha sendmail seem to work fine with the exception > > of very long wait time for mailq (sendmail -bp) - about 25 sec. > > > > I'm not convinced it is an alpha problem, so I post to this list. > > > > On a much slower i386 running the same version of OS and of sendmail > > it takes less than a second. > > > > # uname -srm > > FreeBSD 6.3-STABLE alpha > > > > # sendmail -d0.1 > > Version 8.14.2 > > Compiled with: DNSMAP LOG MAP_REGEX MATCHGECOS MILTER MIME7TO8 MIME8TO7 > > NAMED_BIND NETINET NETINET6 NETUNIX NEWDB NIS PIPELINING SCANF > > TCPWRAPPERS USERDB XDEBUG > > > > The only software difference between my i386 and my alpha I can think of > > is that on alpha I run IPF: > > > > # grep "= 25" /etc/ipf.rules > > pass out on dc0 proto tcp from any to any port = 25 flags S/SA keep state > > Based on your description of the probem, you should ensure that all > block rules in ipf include logging, and see if running mailq is triggering > a timeout because it can't communicate over the network. 25 sec is > suspiciously close to 30 seconds, which is a common timeout value for > network operations. > > sendmail needs to communicate on more ports than just 25, so it's > likely that you've blocked something else that it needs ... DNS perhaps? I solved it first reading http://www.goldfisch.at/knowledge/70 : "If sendmail takes very long time for starting and output of mailq takes many seconds, than you might have a name-resolve-problem. Check logfiles for gethostbyaddr ... failed - entries and fix corresponding resolve-problem." which is what I did: % grep gethostbyaddr /var/log/maillog Feb 15 03:01:23 <myhostname> sendmail[71810]: gethostbyaddr(10.10.10.1) failed: 2 Feb 15 03:01:58 <myhostname> sendmail[71856]: gethostbyaddr(10.10.10.1) failed: 2 Feb 15 03:02:40 <myhostname> sendmail[72061]: gethostbyaddr(10.10.10.1) failed: 2 % I then found why I get this error, from http://www.sunhelp.org/faq/sendmail.html "QUESTION: Why does my Solaris machine generate the error "sendmail: gethostbyaddr failed" whenever I send mail? ANSWER: One of more of the IP addresses configured on your system does not have a matching entry in /etc/hosts. sendmail does a "reverse" lookup (input IP address, output hostname) when it starts up, generating the error. You can either fix /etc/host and your Nameservice, or ignore the message." Although I had an entry for the second ethernet port (10.10.10.1) in /etc/rc.conf, I didn't have it in /etc/hosts. I fixed /etc/hosts and all is fine now. many thanks anton -- Anton Shterenlikht Room 2.6, Queen's Building Mech Eng Dept Bristol University University Walk, Bristol BS8 1TR, UK Tel: +44 (0)117 928 8233 Fax: +44 (0)117 929 4423
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20080215164911.GA45654>