Hoping someone's seen this and/or can help...heh. I'm running the at rpm of dovecot, version dovecot-1.0-0_27.rc10.el4.at. I'm running it on RHEL4, up to date. We're using Maildir (we're upgraded from a Courier-IMAP installation) and things are working...but...we have a few users with very large mailboxes (not their inboxes) where they archive stuff (200k+ messages). When they try to read those mailboxes, the imap server instantly drops the connection, and I get the following in the log:
--cut-- Oct 27 11:20:11 hostname dovecot: IMAP(user): file mail-cache-fields.c: line 26: unreached Oct 27 11:20:11 hostname dovecot: child 27514 (imap) killed with signal 6 --cut--
host and user names changed to protect the guilty. I have tried increasing the imap process size max to 512, as well as doubling the IMAP command size to 131072 (bytes)? Neither made a difference. *other* mailboxes on the same account work fine. The error happens *immediately*, not after a timeout, as soon as they select the mailbox in their client (we're using Evolution, and have tried pine as well...I'm working on trying other clients). The Maildir caching optimization settings in the conf are left to default.
Any ideas? Many thanks.