Aaron Richton wrote:
Is something improperly ordered during teardowns?
[6] ldap_pvt_thread_pool_destroy(tpool = 0x39c374, run_pending = 1), line 570 in "tpool.c"
https://www.nbcs.rutgers.edu/~richton/slapd.log.20090319.bz2
Same issue we ran into earlier, the consumer info has been freed but there's still a consumer task queued in the thread pool. I had already changed things around to avoid this but obviously haven't totally prevented it. Should be fixed now in HEAD.