Regarding a message of yours from a May 2011, I have an 8.3 i386 system where syslogd has been found spinning the CPU a few times in the past week. The stack trace is similar to yours: #0 0x4810cfa9 in _pthread_mutex_init_calloc_cb () from /lib/libc.so.7 #1 0x4810f890 in _pthread_mutex_init_calloc_cb () from /lib/libc.so.7 #2 0x48116bbe in free () from /lib/libc.so.7 #3 0x0804c7e6 in init () #4 <signal handler called> #5 0x48194af3 in select () from /lib/libc.so.7 #6 0x0804e147 in main () On this system, syslogd is configured with a pipe to sshguard to monitor failed logins. In my searches, I have not noticed any resolution for this issue -- any chance I've missed one? Thanks, Guy