FYI, we saw one panic on our director/lmtp-proxy yesterday. It's running dovecot v2.0.13, so it very well might have been fixed already.., but here it is anyway:

Aug 31 11:33:31 loadbalancer1 dovecot:: lmtp(4119): Panic: file lmtp-proxy.c: line 370 (lmtp_proxy_output_timeout): assertion failed: (proxy->data_input->eof) Aug 31 11:33:31 loadbalancer1 dovecot:: lmtp(4119): Error: Raw backtrace: /usr/lib64/dovecot/libdovecot.so.0 [0x3aabc3bb70] -> /usr/lib64/dovecot/libdovecot.so.0 [0x3aabc3bbc6] -> /usr/lib64/dovecot/libdovecot.so.0 [0x3aabc3b083] -> dovecot/lmtp [0x406c90] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_handle_timeouts+0xcd) [0x3aabc46a8d] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_handler_run+0x68) [0x3aabc47a48] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_run+0x2d) [0x3aabc466fd] -> /usr/lib64/dovecot/libdovecot.so.0(master_service_run+0x13) [0x3aabc34fe3] -> dovecot/lmtp(main+0x144) [0x404034] -> /lib64/libc.so.6(__libc_start_main+0xf4) [0x32ae81d994] -> dovecot/lmtp [0x403e39] Aug 31 11:33:31 loadbalancer1 dovecot:: master: Error: service(lmtp): child 4119 killed with signal 6 (core dumps disabled)

On postfix side at the same time, we saw a few different messages about:

postfix/lmtp[5178]::  849FD1C024: lost connection with loadbalancers.example.com[192.168.42.15] while sending end of data -- message may be sent more than once

so I suspect this crash took down all active lmtp-proxyed connections.

-jf