https://bugs.openldap.org/show_bug.cgi?id=8901

--- Comment #7 from Howard Chu hyc@openldap.org --- (In reply to Marcelo.DeCastroLoebens from comment #6)

Thanks for the analysis and workaround provided.

Do you have an opinion about the risk associated with the workaround? Since I'm not familiar with the code, that would be very helpful.

I don't have the expertise to comprehensive test slapd in isolation, so I would like to know if you imagine possible side effects during "regular" usage (not shutdown).

There's no impact during regular use since ldap_pvt_thread_pool_close() is only called during shutdown.

Actually, just to be more explicit - the sleep(1) should be inserted in slapd/daemon.c:slapd_daemon_task() right after its call to ldap_pvt_thread_pool_close(). There's no need to muck with libldap itself.