When my libdb was upgraded from 4.1 to 4.2. This makes the database-files incompatible: /etc/init.d/slapd start goes ok (but I can see in /var/log/messages that slapd fails to load the database and dies) /etc/init.d/slapd stop then fails, since it can't find the process listed in the pid file, and you have to do a /etc/inid.d/slapd zap if you want to try to start it again. So to add things up: When upgrading berkely db from 4.1 to 4.2, you need to rebuild database files (index, cache, log) slapd initscript fails to detect if slapd dies fast (it can sleep for 250 milliseconds and see if /proc/PID/status exists or something like that) Reproducible: Always Steps to Reproduce: upgrade berkelydb after slapd has been setup Actual Results: things look okey during boot Expected Results: init script should fail, and post a message what might be the cause
This is fixed in latest baselayout You can adjust the time checking for a process being up there
