Robin Green wrote:
I would suggest you just make you of the way init works and chkconfig work. TryFedora-list,
I just rebooted my server and openldap failed to start, which broke email; this is because my openldap installation depends on a postgresql database and postgresql starts later in the bootup sequence.
I realise that there isn't always "one right solution" for everyone when it comes to bootup order. I am storing my ldap database in postgresql; other people might use ldap to authenticate to postgresql, so they would require the current startup order.
So, I fixed it for now (at least, in runlevel 3 - runlevel 5 is never used on this server) by:
cd /etc/rc3.d mv S64postgresql S38postgresql
But might not this get broken the next time I upgrade postgresql? And if so, is there anything I can do about it - short of maintaining my own custom postgresql rpm and prioritising my own yum repository highest in yum.repos.d?
#chkconfig --del openldap
then edit the your openldap startup script, which i'd presume to be /etc/init.d/openldap. Change the line which reads "chkconfig: 2345 08 92" to something more appropirate for you. From what I remember, "2345" represent the runlevels on which it gets started. "08" represents the start number. and "92" represents the stop number. Then
#chkconfig --add openldap
I would suggest you backup the init script first. And if things work out, then just save a copy of the edited init script and restore it when ever openldap gets updated.
Just a suggestion.