Hi, On a new server, I was hit by a 'signaled to death by 11' problem recently. Logs looked like this:
Jan 15 00:11:27 xxxx imapd[13435]: open: user xxeid opened INBOX Jan 15 00:11:27 xxxx master[1838]: process 13435 exited, signaled to death by 11 Jan 15 00:11:27 xxxx imapd[13617]: accepted connection Jan 15 00:11:27 xxxx imapd[13617]: login: [192.168.9.2] xxeid plaintext Jan 15 00:11:27 xxxx imapd[13617]: seen_db: user xxeid opened /var/lib/imap/user/x/xxeid.seen Jan 15 00:11:28 xxxx imapd[13617]: open: user xxeid opened INBOX Jan 15 00:11:28 xxxx master[1838]: process 13617 exited, signaled to death by 11 I checked the archives and found similar problem reports and the solution was to not use pam_ldap for authentication. I have then rebuilt cyrus-sasl with it's own, native LDAP support. Unfortunately the 'signaled to death by 11' problem didn't go away. Somebody pointed me to a description of another 'signaled to death by 11' problem where the solution was to reconstruct the mailbox. I tried reconstruct -r user.xxeid and the problme was solved. Now, my question is: - Why does this problem only affect about ~3% of the users on this server. - All mailboxes were empty. They have been created by a script which periodically reads the LDAP tree and creates mailboxes as needed. - Should I call reconstruct -r for every newly created mailbox? Any feedback is welcome Simon The server config is: name : Cyrus IMAPD version : v2.1.11-Invoca-RPM-2.1.11-6 2002/12/04 14:53:12 vendor : Project Cyrus support-url: http://asg.web.cmu.edu/cyrus os : Linux os-version : 2.4.18-18.7.xsmp environment: Cyrus SASL 2.1.10 Sleepycat Software: Berkeley DB 3.3.11: (July 12, 2001) OpenSSL 0.9.6b [engine] 9 Jul 2001 CMU Sieve 2.2 TCP Wrappers mmap = shared lock = fcntl nonblock = fcntl auth = unix idle = poll mboxlist.db = skiplist subs.db = flat seen.db = skiplist duplicate.db = db3-nosync tls.db = db3-nosync
