I am running Cyrus 2.2.3 (Simon's RPMS) and sendmail 8.12.10 with the socketmap
patch. Cyrus is running in a murder configuration, with virtdomain support and
ldap (via saslauthd).
Hardware wise, we are currently using 2 frontend servers (Dell Poweredge 1750's
2.8ghz Xeons and 1 gig of ram), and 1 backend server (a Dell Poweredge 2650, I
forget the specs). The OS's are Fedora Core 2.
We recently started having issues with sendmail would "hang" for a period of
time. This would cause the number of sendmail processes running to rise (in
return causing the open connections to the ldap servers to rise). In trying to
figure out what is causing this, I ran an strace on a sendmail process that had
been sitting in the process list for some time. This is what I saw:
Process 20449 attached - interrupt to quit
connect(7, {sa_family=AF_FILE, path="/var/lib/imap/socket/socketmap"}, 110) = 0
And it just sits there. Almost as if it's unable to connect to the socket. At
this point in time there were approximately 250 sendmail processes running (and
rising, because they all get "stuck" trying to access the smmapd socket).
This has me really stumped. This seems to be the only thing causing bad access
times for the whole system. Does anyone have any idea?
Thanks.
Lenny
--
"Wisdom is to a man an infinite Treasure" - Anonymous"
---
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
